You are on page 1of 47

Annual Report 2003

Power Systems and

High Voltage Laboratory Issued by

Power Systems and High Voltage Laboratory

(Institut für Elektrische Energieübertragung und Hochspannungstechnologie)
Annual Report Swiss Federal Institute of Technology (ETH) Zürich
2003 ETH Zentrum, Physikstrasse 3, CH-8092 Zürich

Power Systems Laboratory High Voltage Laboratory

Phone: +41 – 1 –632 41 86 Phone: +41 – 1 –632 27 77

Fax: +41 – 1 –632 12 52 Fax: +41 – 1 –632 12 02
Email: Email:

Front cover: Conceptual view of a power system with Wide Area Monitoring
and Control. See page 20

Back cover: Electrical field simulation in a two-dimensional representation of

the switching chamber of a generator circuit breaker. Field
strength is depicted by color, where blue corresponds to low
strength and red high strength. The simulation is designed to
gather information relevant to the placement of electric field
probes exterior to the breaker chamber.

eeh elektrische energieübertragung

und hochspannungstechnik
Dear Friends of the Laboratory,

The size of the research staff has during the last year stabilized at a level of about
30 active researchers, among which around 25 are PhD students. This seems to be
an appropriate size with regard to available resources in form of office space,
senior researchers, and funding. As seen from the detailed descriptions of the
research projects in this annual report, many of the projects are carried out in
collaboration with or with support from industrial partners. We are grateful for
the confidence in the work performed at the laboratory shown by our external
sponsors. Members of the staff have also been very active in different
international conferences, workshops, and seminars by presenting papers and
results based on the research results from various projects. Many of the papers
and reports can be found at our homepage,, where the
progress of the different research projects is continuously updated.

We are particularly pleased that the project concerning Future Energy Networks
has gained additional momentum. At the moment four PhD students, two from
the power systems laboratory and two from the high voltage laboratory, are
involved in the project, and two industrial partners, ABB and Areva (former
Alstom), are sponsoring the project. Discussions are going on with additional
external partners. We also hope that our colleagues from the power electronics
laboratory and automatic control laboratory could join us with collaborators
during the next year. It is our intention and hope that we through this project can
contribute to improving the efficiency and sustainability of the energy supply in
the future, which we believe will be one of the most important questions in the
coming decades. More about this project can be found on page 85.

The spin-off company Plexim, which was formed by two members of the power
systems laboratory two years ago and has its office in direct connection with the
power systems laboratory, has been successful in promoting its products. During
the year a project involving Plexim, ABB, and the power systems laboratory, was
approved by KTI, Swiss Federal Commission for Technology and Innovation. This
new project aims at further developing the products offered by Plexim, see page
22. We wish the founders of Plexim, Drs J. Allmelling and W. Hammer, all the best
with their company.

During the year Frau B. Rutz retired as secretary of the high voltage group after
many years in this capacity. We want to thank Frau Rutz for excellent services and
her positive attitude, and we hope that she will enjoy her retirement. We also
want to welcome Frau K. Sonderegger as new secretary in the group. It is our hope
that she will find our laboratory a stimulating and interesting place to work.

Least but last we want to thank all the personnel of the laboratory for their
enthusiastic and excellent work performance. The most important asset of an

academic institution is its staff, and thanks to the good ideas and hard work of all
our collaborators we can conclude that 2003 was another successful year of the Contents




G. Andersson K. Fröhlich
1. Organisation...................................................................................................... 1
2. Teaching.............................................................................................................3
2.1 Lectures 3
2.2 Seminars 5
2.3 Student Projects 5
2.4 Diploma Projects 5
2.5 Excursions 6
3. Research Activities............................................................................................7
3.1 Completed PhD Theses 7
3.2 Current Projects 11
3.3 Spin-Off Activities 22
4. Publications and Reports...............................................................................23
4.1 Publications 23
4.2 Reports 24
5. Presentations ..................................................................................................25
6. Conferences and Visits ..................................................................................27
6.1 Conferences and Workshops 27
6.2 Visits 29
6.3 Awards 29


1. Organisation.................................................................................................... 31
2. Teaching...........................................................................................................33
2.1 Lectures 33
2.2 Student Projects 35
2.3 Diploma Projects 35
2.4 Excursions 36
3. Research Activities..........................................................................................37
3.1 Completed PhD Theses 37
3.2 Current projects 41
3.3 Services offered 74
4. Publications and Reports...............................................................................75
4.1 Publications 75

iv v
5. Presentations ..................................................................................................77
6. Conferences and Visits ................................................................................. 79
Activities of the Power Systems Laboratory
6.1 Conferences and Workshops 79
6.2 Events 81
6.3 Visits 81 1. Organisation
JOINT ACTIVITIES 83 Head: Prof. Dr. Göran Andersson
1. Colloquia ......................................................................................................... 83 Secretary: Rita Zerjeski
2. Vision 'Future Energy Networks'................................................................. 85
Scientific Staff: Dr. sc. techn. Jost Allmeling
Dipl. El.-Ing. Martin Geidl from 1 July 2003
Dipl. El.-Ing. Wolfgang Hammer
Dipl.-Ing. Andrei Karpatchev until 15 October 2003
Dipl. El.-Ing. ETH Gaudenz Koeppel
Dipl. Wirtsch.-Ing. Thilo Krause
Mirjana Milosevic, M.Sc. El.Eng
Derek Poon, cand. B.El.Eng from 15 May 2003
Rusejla Sadikovic, M.Sc. El.Eng
Dipl. El.-Ing. ETH Christian Schaffner
Marek Zima, M.Sc. El.Eng

Scientific Associates: Prof. em. Dr. Hans Glavitsch

External Lecturers: Dr. Rainer Bacher, Bundesamt für Energie, Bern

Dr. Dieter Reichelt, Technische Betriebe, Kreuzlingen
Dr. Christian Rehtanz, ABB Power Technologies Management Ltd


2. Teaching
The lectures and laboratory classes listed in the following section are part of the
curriculum of the Electrical Engineering Department and are conducted by the staff of
the Power Systems Laboratory. Details of the entire electrical engineering curriculum
can be provided on application (list of lectures, option proposals).

2.1 Lectures

5th semester 4G
Electric Power Systems Andersson, G.
Elektrische Energiesysteme Fröhlich, K.
Structure of electric power systems; symmetric three phase systems; modelling of
power transformers, generators and transmission lines; analysis of symmetrical and
unsymmetrical three phase systems; transient switching phenomena; basics of current
interruption; principles and applications of distribution- and transmission switchgear;
basics of insulation coordination.

6th semester 4G
Modelling and Analysis of Power Networks Andersson, G.
Modellierung und Analyse elektrischer Netze
The electrical power transmission system, the network control system, requirements
for power transmission systems (supply, operation, economics), network planning and
operation management, models of N-port components (transmission line, cable, shunt,
transformer), data specification per unit (p.u.), Linear Modelling of networks, Linear und
non-linear calculation (NewtonRaphson), non-linear load flow (specification and
solution methods), threephase und generalized short circuit current calculation, further
applications of load flow calculation. Introduction to dynamics and stability in power
systems. Rotor angle and voltage stability. Equal area criterion. Control of power

7th semester 4G
Optimization of Liberalized Electric Power Systems Bacher, R.
Optimierung liberalisierter elektrischer Energiesysteme
Mathematical optimization methods, Karush-Kuhn-Tucker optimality conditions,
Equality constrained non-linear optimization, Linear Programming (LP) (Simplex,
Interior Point), Quadratic Programming (QP) and applications; Non-linear optimization;
Goals of a power exchange (PX), of the "independent system operator" (ISO), of the
regulatory institutions, of the new electric power utilities; Principles of optimization of
a power exchange: Offer-Bids; Optimal regulated network operation: Payments for
network use, long- and medium term network optimization as goals; Handling of
network security limitations by optimization methods; Optimization methods to
determine the efficiency of networks; Optimization of ancillary services as part of the
liberalized electric power system.

2 3

7th semester 2G 2.2 Seminars

Portfolio and Risk Management in a Liberalized Market 1 D. Reichelt
Portfolio und Risk Management im liberalisierten Strommarkt 1
1st-4th semester 4 PPS
Open power markets all over the world, market models of the European countries, PPS: Economical and technical aspects of a sustainable energy supply G. Koeppel
Swiss electricity market today, pan-european power trading, management of the Wirtsch. und techn. Aspekte einer nachhaltigen Energieversorgung Th. Krause
physical portfolio of power plants and delivery contracts, power market indices,
hedging strategies using futures, case study (1): hedging strategie, European energy In the past, electricity markets were characterized by vertically integrated utilities
exchange (EEX), energy market risk (value at risk, profit at risk), options and structured operating as regulated monopolies. However, the ongoing liberalisation process, the
products in the power market, enterprisewide risk management (Basel II), case study Kyoto-protocol as well as upcoming technologies are forcing a reorganisation and
(2): Barings Bank, introduction of swaps and other derivatives on electricity prices. redirection of the electricity market.
The offered seminar addresses several issues related to this reorganisation process.
Main topics are distributed generation, particularly aspects of renewable energy
8th semester 2G sources (solar and wind power) as well as economical and ecological issues on
Portfolio and Risk Management in a Liberalized Market 2 D. Reichelt liberalized markets. The students are writing and presenting a report covering single
Portfolio und Risk Management im liberalisierten Strommarkt 2 aspects, learning how to search for literature as well as how to write and present
Understand the worldwide liberalisation of electricity markets, the various markets scientific reports.
models, pan-european power trading and the role of power exchanges. Financial
products (derivatives) based on power prices, management of a portfolio containing
physical production, contracts and derivatives, evaluation of trading and hedging 2.3 Student Projects
strategies, methods and tools of risk management.
To be admitted to the diploma examinations of the 7th and 8th semester, students of
8th semester 4G the electrical engineering department must carry out two projects. Each student can
Power System Dynamics and Control Andersson, G. freely choose his subject area, but usually the two projects have to originate from
Systemdynamik und Leittechnik in der el. Energieversorung Rehtanz, Ch. different subject areas. According to the curriculum, two days of the week during the
semester period are to be devoted to this work. In general, the subjects are derived
Dynamic properties of electrical machines, networks, loads and interconnected from topical research and development tasks. As we have close collaboration with the
systems. Models of power stations and turbines, control of turbines, load- and FKH some of the student projects as well as diploma work (see the following section)
frequency control, power exchange between networks, model of the synchronous are supervised by staff of the FKH.
machine connected with the network, behaviour of the machine in case of
disturbances, transient stability, equal area criterion, model for small disturbances, G. Glanzmann, Implementation of a Market Model in an Interactive Power
voltage control. Facts-Devices. SCADA/State Estimation. EMS-Implementations, M. von Siebenthal Flow Simulation Platform
Protection, Asset Management, Future Trends in IT for Power Systems.
R. Wohlwend Beschleunigte Berechnung des stationären Zustandes
leistungselektronischer Systeme"

2.4 Diploma Projects

Allocated time is four months. The majority of students devote their time to this work
in the winter semester. The student has the option to carry it out either before or after
the formal diploma examination (dates in spring and autumn).

Michele Bernocchi Simulation of Transmission Pricing Methods for Liberalized

Thomas Tarnowski FACTS Devices Optimizing an Electrical Energy Market subject to
Power Transmission Constraints

4 5

2.5 Excursions 3. Research Activities

Credit Suisse First Boston, Zürich; 10 June 2003 3.1 Completed PhD Theses
Trading Floor visit and presentation about structured financial
Candidate: Dipl. Ing. Wolfgang Hammer
NOK, Baden; 17 June 2003 Thesis: ETH No. 15269
Date of oral examination: 19 September 2003
Examiner: Prof. Dr. G. Andersson, ETH Zürich
Co-examiner: Prof. Dr. A. Golé, University of Manitoba, Canada

Author's summary
For more than one hundred years, the generation, transmission, distribution and
utilization of electric energy has been based principally on alternating current.
However, there are a number of special applications for which direct current
transmission is the better if not only choice, even taking into account the cost of the
equipment that is necessary to convert between ac and dc:
• Bulk power transmission on long overhead lines At very long distances, ac
overhead lines consume large amounts of reactive power which are furthermore
dependent on the amount of active power transfered. In addition to the additional
losses caused by the reactive current, this also gives rise to stability problems. DC
lines, on the other hand, do not consume reactive power.
• Power transmission via cable Due to its construction, a cable has a much higher
capacitance per unit length than an overhead line. This means that even for a cable
of moderate length (50 km), the reactive current can utilize a major part of the total
current capability when transmitting ac power.
• Transmission between unsynchronized ac systems AC power transmission is
physically only possible between synchronized systems. When two systems operate
at different frequencies (such as 50 Hz and 60 Hz), the only practical way to
transmit power between them is by means of a dc connection.

In addition to these more traditional applications concerning rather the stationary

operation of networks, there is a fourth aspect of HVDC transmission which gains more
and more importance with the growing global interconnection of transmission
• Parallel ac and dc transmission In interconnected ac systems technical problems
like transient instability or power oscillations can occur when an interconnection
between two parts is relatively weak. In such a case, the fast controls of an
additional HVDC link in parallel to the ac interconnection can be used to stabilize
the system.

With today's power systems being operated closer to their stability limits, and
particularly in view of this last aspect, there is an increasing need to understand the
dynamic interactions between HVDC converters and the ac system. Convenient models

6 7

are needed that will both facilitate control design and allow fast and accurate transient INCREASED TRANSMISSION CAPACITY BY FORCED SYMMETRIZATION
simulations. However, the analysis of HVDC converter systems is challenging because
of their hybrid nature, as they incorporate both continuous-time dynamics (associated Candidate: Dipl. Ing. Andrei Karpatchev
with the voltages and currents of capacitors and inductors) and discrete events (due to Thesis: ETH No. 15342
the switching of the valves). Date of oral examination: 5 December 2003
Examiner: Prof. Dr. G. Andersson, ETH Zürich
Transient stability programs typically use power electronics device models that are Co-examiner: Prof. em. Dr. H. Glavitsch, ETH Zürich
based on quasi-static approximations. Such models rely on the assumption that the ac Prof. em. Dr. D. Povh, University of Ljubljana
system is in sinusoidal steady-state and, in case of a current source HVDC converter,
that the direct current is ripple-free and constant. Due to these assumptions, the Author's Summary
models are strictly valid only in the steady state and lack accuracy when transient The electrical power industry experiences nowadays a significant need in modern
stability or other dynamic phenomena are of interest. techniques for increasing the capability of power transmission systems. The reason for
this lies in growing power flows, caused by rising power consumption, and in
This work presents a new dynamic modeling approach for two types of HVDC deregulation of the electrical market, where power flows should be more flexible. These
converters: the conventional Line Commutated Converter (LCC) and the Capacitor factors demand more transmission capability from the existing networks. The
Commutated Converter (CCC). This approach takes into account the dynamics of a conservative expansion of the high voltage grid is often not desirable or not possible,
variable direct current and (for the CCC) of the series capacitor voltages. The because the approval of new overhead transmission lines meets often strong
investigations are restricted to phasor-based models for use with transient stability opposition. Furthermore, it takes a long time and is generally a risky long-term financial
programs. The resulting dynamic converter models are natural extensions of the investment. New technological solutions are sought to meet the capability needs under
standard quasi-static models and are in fact identical in the steady state. consideration of modern environmental requirements.
Case studies demonstrate the impact of the new models on the study of control The presented work considers possibilities for ensuring power transmission through a
interactions between the pole controller of an HVDC converter and the ac system, an AC transmission line with a damaged phase conductor. The disturbance of the
application normally considered outside the scope of phasor-based simulations. For the symmetrical pattern of currents and voltages in the surrounding network can be
LCC, there is only a marginal difference between the quasi-static and the dynamic eliminated by active measures for symmetry. The utilization of two remaining healthy
model. For the CCC, on the other hand, the simulations clearly reveal the superiority of phases of a three-phase transmission line with a damaged phase in this way can be an
the proposed dynamic model over the quasi-static model. Two cases are presented in economical way to enhance the system reliability.
which the static model leads to a wrong estimation of the system stability, whereas the
response of the dynamic model matches a detailed simulation nearly perfectly. The present planning procedure of the high voltage networks mostly respects the (n-1)
criterion. This means that the network should not be subject to any overload or voltage
drop below a strictly given limit when any network element is disconnected. Based on
statistics, single phase-to-ground faults are the most frequent faults in transmission
systems, typically approximately 2/3 of all line faults at voltage level 220-380 kV.
Figure 1: Present planning procedures are often based on single outages of three-phase circuits,
Response to a current which do not take into account the actual fault pattern. For the single-phase faults it is
order step change from necessary to avoid unsymmetrical conditions or unsymmetrical currents in the
1.8 kA to 2.0 kA at 0.02 s: network. The reason for this is that the currents in the zero sequence system mean
(a) ac bus phase voltage, earth currents, and those can be dangerous for people and cause adverse interactions
(b) direct current, with other systems. The currents and voltages in the negative-sequence system are of
(c) firing angle. concern to rotating synchronous machines like generators and motors, but if no such
Thin solid line: detailed machines are connected to a network part, the negative-sequence voltages can be
simulation; heavy solid tolerated in that part.
line: dynamic model;
dashed line: standard Symmetrization means the suppression of both zero- and negative-sequence currents
on the network side of both line circuit breakers so that the network does not
experience any unsymmetrical conditions. This can be performed by the installation of
special equipment as shunt or serial elements at the line terminals.

8 9

Different arrangements and strategies are considered in this work. In order to be able 3.2 Current Projects
to try all the different arrangements a special system simulator has been developed. It
is based on power flow calculations with multiple symmetrical system representations
and allows simulation of the symmetrization effect in a complex meshed network. The PROTECTION AND CONTROL SCHEMES FOR NETWORKS WITH A LARGE
fault currents in the negative and zero sequence systems can be studied directly. CONTINGENT OF DISTRIBUTED GENERATION
Different symmetrization topologies, both concentrated and distributed, are
investigated and discussed. Martin Geidl

Modern power electronics devices and measurement technology provide the hardware Introduction
basis for the practical implementation of different modern symmetrization techniques. The global trends of reducing greenhouse gas emissions and liberalising utility
The present work considers impact on the whole grid, in case symmetrization industries lead to an increasing utilisation of small, often private owned and mostly
techniques are applied. The achieved rest transmission capacity of the damaged line renewable power sources, which are commonly identified as 'distributed', 'embedded'
and other "system" characteristics are in focus of interest. Different methods of or 'decentralised' generation (DG). Beside various benefits, several technical,
symmetrization are considered and compared. A possibility of distributed economical and regulatory questions come up when integrating these sources in
compensation is also one item of interest, which is very promising for the protection of existing networks.
several lines. In terms of physical integration, protection is one of the major issues. Utility relaying
often fails due to changed short circuit levels, line impedances, reverse power flow etc.
Symmetrization methods, which are examined in this work, considerably increase the Therefore new schemes for operating the grid in case of a failure have to be developed,
system reliability and can be seen as competitive solutions for an extensive grid whereas both protection and control have to response properly.
expansion on the existing congestion routes. The installation of the necessary An advanced system combines the functionality of protection (classical as well as wide
additional equipment can be done in shorter terms than the construction of a new area) and control, the border in between gets unclear resp. disappears. Modern
transmission line. Environmental and aesthetic impact is then very small too. The information technology has to be used to comply with the ambitious communication
modern power electronics devices, which are needed to achieve these advantages, requirements.
introduce more operational flexibility and can be used also in non-disturbed operation Goal of the project
of the network. The aim of this project is to develop protection and control philosophies for future
networks with a large contingent of distributed generation. Additionally, the
communication requirements of an integrated protection and control system should
be estimated. The developed strategies will be verified using typical grid topologies
which are preliminarily defined.

This research is performed within the framework of the project Vision Future Energy
Networks (please refer to the section 'Joint Activities').

[1] Dugan, R. C.; McDermot, T. E.: "Operating Conflicts for Distributed Generation on
Distribution Systems", IEEE Rural Electric Power Conference, 2001.
[2] Dondi, P. et al: "Network integration of distributed power generation", Journal of
Power Sources, 2002.
[3] Wallace, A. R.: "Protection of embedded generation schemes", IEE Colloquium on
Protection and Connection of Renewable Energy Systems, 1999.
[4] Böse, C. et al: "Protection System Coordination - Considering New Techniques by
Using Flexible Models", 14th Power Systems Computation Conference, 2002.

Partnerships: ABB Switzerland, Alstom France, ETH Zürich

10 11

DEVELOPMENT OF METHODS FOR THE OPTIMAL SIZING AND POSITIONING OF defining the requirements for storage in terms of energy and power over time. These
requirements then serve as input to more technologically oriented research areas
STORAGE DEVICES within the project.
Gaudenz Koeppel Objective
Introduction The intention is to scientifically contribute to the discussion on the needs of storage
The increasing number of small-scale generators, often producing electricity from devices in future energy networks. This discussion ranges from views, implying that
renewable sources, embedded in the sub-transmission and distribution networks, leads information technology allows to perfectly match production and consumption, thus
to several new issues and effects in the infrastructure and behaviour of power supplies. turning storage redundant to views, suggesting that every stochastic generator needs
There are e.g. increased power fluctuations on the grid due to the stochastic nature of to be combined with a storage device. The corresponding task is consequently to find
renewably produced electricity or e.g. different protection requirements, because of out, where how much storage will be needed, to fulfil given supply security
limited fault-ride-through capabilities (a literature review of this topic is given in [3]). requirements.
Furthermore some very large cities begin to be confronted with capacity bottlenecks
during peak load times, having limited options for building more transmission lines.
[1] Billinton, R., Allan, R.N.: "Reliability Evaluation of Power Systems", 2nd Edition,
Storage devices could be a solution to some of these effects and it is the goal of this Plenum Press, New York 1996
project to develop methods for the optimal sizing and positioning of a storage device [2] Kerstin, W.H.: "Distribution System Modelling and Analysis", Electric Power
for a given infrastructure. Depending on the grid topology possibly including local Engineering Series, CRC Press, Boca Raton 2002
generation, the consumer structure and behaviour as well as requirements for [3] Koeppel, G.: "Distributed Generation - Literature Review and Outline of the Swiss
availability, reliability and power quality, the optimal size (energy and power) and place Situation", Technical Report, Zurich 2003; available online:
for a storage device should be determined.

Partnerships: ABB Switzerland, Alstom France, ETH Zürich


Thilo Krause
That a country’s electricity market has been liberalized is a common but likewise
unspecific proposition. "Delivered power is a bundle of many services. These include
transmission, distribution, frequency control and voltage support, as well as
generation. [...] Each service requires a separate market, and some require several
Whereas the theory for each single market is nowadays well developed, there is still a
need for studying the interaction of the different sub-markets. This especially applies to
the energy-, transmission- and ancillary services markets. Generators may bid on a pool
Figure 1: Symbolic illustration of the impact from a possible placement of a storage device (flywheel) for
given load profiles and local power generation (wind turbine and fuel cell) (or spot) market or directly settle into bilateral contracts with particular customers. This
energy trade results in a physical power flow through the transmission network, where
Figure 1 shows a symbolic example of how the cumulated load profiles for a sample the grid should be capable of transmitting the energy according to the agreed set of
network could look like after inserting a storage device. As the figure implies, the contracts. In order do so transmission pricing methodologies have to be developed to:
storage device could basically be placed anywhere, e.g. either to momentarily balance "a) promote economic efficiency b) compensate grid companies fairly for providing the
the stochastic output of the wind turbine or to e.g. balance the day-night differences of transmission service c) allocate transmission cost reasonably among all transmission
a part of the investigated grid – it depends only on the existing requirements. users d) maintain the reliability of the grid."[2] Methodologies reach from embedded to
This research is performed within the framework of the project Vision Future Energy marginal pricing schemes. For the ongoing project it is proposed to use embedded
Networks (please refer to the section 'Joint Activities') and contributes to this project by methods for cost recovery and marginal methods for setting incentives to promote the

12 13

efficient use of the network in the short-run and to ensure reliability in the long-run. voltage of varying frequency to an ac voltage of system frequency. The requirements on
The computed transmission prices serve as additional input data for the energy market, the permissible deviation within 0.1 Hz, or even smaller values are accepted in power
influencing the strategic behaviour of generators and loads. Thus, demand and supply systems used in the industrialized countries. In smaller systems, like the ones studied in
are a function of generation- as well as transmission and reliability constraints. this project, much larger frequency deviations must probably be accepted for an
economic operation. This requires higher demands on the converters between the
Objective energy sources and the network.
A report evaluating different transmission pricing methodologies is available since Thus, it means that models for the interaction between the network side and the
November 2003[3]. The report serves as initial study for the overall project. The energy sides of the converters are needed. If several of these are present in proximity of
objective for the next research stage is to implement the studied methods for the each other, interactions between them could arise. One application of these models is
Deregulated Market Simulator (DemSi). The modelling of the energy market is carried the specification of the design criteria of the power electronic equipment and for
at the EPFL Lausanne. Both parts will be merged presumably in summer 2004 to form a determining the electrical conditions at the energy source. Another application is for
first prototype of the simulator. The simulator will be extended stepwise and used for studying the interaction among several converters in a network and determining the
case studies in order to verify the chosen approach. possible adverse condition that could arise.
So far, studies of the interaction between hysteresis controlled voltage source inverters
connected to the same power network are analyzed [1].
[1] Stoft, S. (2002): "Power System Economics. Designing Markets for Electricity.", The coupling between the inverters results in an interdependence of their switching.
Piscataway 2002. The effects of various parameters of the inverters are analyzed. One of the aims of this
[2] Cannella, M.; Disher, E.; Gagliardi, R. (1996): "Beyond the Contract Path: A Realistic project is to develop models for the applications above.
Approach to Transmission Pricing", The Electricity Journal. 1996. pp. 26 – 31.
[3] Krause, Th. (2003): "Evaluation of Transmission Pricing Methods for Liberalized The following activities of this project are:
Markets - A Literature Survey", Technical Report, Zurich, 2003; available online: 1) Development of models of the interface between the energy sources and the network
2) Studies of the stresses on the power electric equipment in typical model networks
Partnerships: PSEL, ETH Zürich
[1] M. Milosevic, J. Allmeling, G. Andersson, "Interaction between Hysteresis
ISOLATED RURAL DISTRIBUTION NETWORKS WITH A LARGE PENETRATION OF Controlled Inverters used in Distributed Generation Systems", submitted for
presentation at IEEE Power Engineering Society General Meeting, June 2004,
Mirjana Milosevic
Partnerships: Alliance for Global Sustainability, AGS
Participating Institutions
Chalmers University of Technology, Sweden
MIT, Cambridge, MA, USA
Shanghai Jiao Tong University, China Rusejla Sadikovic
Goal of the project Goal of the project
The principal scientific aim of this project is the development of methods for assessing The overall goal of this project is to develop an entire control system architecture,
the performance of rural electricity distribution networks with a large penetration of which avoids complete system redesign in order to enable a most effective system
renewable sources of energy (like wind, sun and flow-of-river). The research expansion in terms of new transmission capabilities and more effective network
concentrates on the development of models for reliability analysis of the supply, but utilization. For the realization of such a controller design the following specifications
some of the models can also be used in voltage quality assessment studies. Additional are defined:
results of the project are expected on the development of techniques for improving the
• No adverse interactions occur when the controllability is lost due to faults or
performance and/or reducing the cost.
changes in system topology.
Abstract • No adverse interactions occur in normal operation, and when normal changes in
The integration of energy sources requires an interface to the electrical system. The control set points are introduced.
interface usually consists of a converter system converting a DC voltage or an ac

14 15

Abstract the fact that such a device can be built in such a way that it can be relocated due to
In the traditional expansion of a planning system, or the replacement of an existing seasonal differences in the power transfer between regions.
main component, extensive studies and redesign of control and protection systems are
needed to guarantee an acceptable systems behaviour concerning performance and Objective
reliability. The concept of Non Intrusive Control Systems (NISC) has been introduced to In this PhD thesis it will be examined what additional economic value the increased
overcome this problem. The basic idea in this concept is to apply a control strategy that transmission capability provided by controllable devices (such as FACTS) gives in a
enables the operation of a controlled transmission path without affecting the rest of liberalized electrical power transmission system.
the system. That control architecture is called Non-Intrusive System Control (NISCTM). Outline
NISCTM avoids complete system redesign. It enables a most effective system expansion To achieve the objectives described in the preceding paragraph several research areas
and more effective network utilization by considering the needed transmission in the field of electric power transmission and controllable devices in such systems are
functions first. The goal of the NISCTM architecture is to simplify the design process so examined.
that new controlled transmission paths can be designed without extensive system The idea behind our model is that a complicated algorithm using optimal power flow
studies. For the operation of a new transmission path the NISCTM architecture avoids (OPF) is often too complex for the purpose of valuating such projects: Input parameters
adverse control interactions within the entire system without causing a redesign of such as expected future generation mix are often hard to guess and error prone.
already implemented controllers. Therefore a high degree of robustness and an Therefore it is reasonable to use a simplified model of the underlying electricity system.
effective design procedure can be achieved. Additionally, the proposed architecture
allows for a proper reaction on critical events and avoids insufficient and hence wrong This method is based on a model that uses area-constraints and aggregated generation
operation after the power system state changes. in each area coupled with a double auction market model, based on the production
costs of generators.
[1] "Non-intrusive control system architecture for ac power transmission", AC-DC To perform the actual valuation of the FACTS one needs to find a framework suitable
Power Transmission, 2001. Seventh International Conference on (Conf. Publ. No. for this special kind of project. Here a so-called Special Purpose Vehicle company (S.P.V)
485), 2001. is used: This company is a separate legal entity from the power producer but is -- in our
study -- fully owned by the power producer. This is necessary to comply with the
Partnerships: ABB Switzerland regulation that generating companies need to be financially separated from companies
working in the field of electricity transmission. The S.P.V borrows the full amount it
needs to fund the project from the power producer. It then buys electric energy in the
low price zone and re-sells it in the high price zone. The cash flow is in the opposite
VALUE OF CONTROLLABLE DEVICES IN A LIBERALIZED ELECTRICITY MARKET direction: it gets paid from the consumers in the high price zone and gives back the
Christian Schaffner profit to the power producer company (see Figure 1).

In a liberalized electricity market, the transmission capability of an electrical network is Power Producer Ltd. Power Power Customer
of economic value to the network operator, which usually is regarded as a natural Sale 100% owned by
in low price zone in high price zone
monopoly, combined with the mission to maximize the benefit for its users while power producer
Profit Payment
giving a reasonable profit to its owners. Due to physical constraints, lines are often only
utilized at a fraction of their physical limits. To improve customer benefit one Figure 1: A Special Purpose Vehicle (SPV) is used as business model
possibility would be to raise the economic value of the transmission lines by increasing
the power transfer capability of these lines. Additionally, there will be a gain in overall In different case studies the influence of various parameters on the value of the
market efficiency since more energy trading can take place between competing regions controllable device is determined. Typical parameters are: Investment costs, running
with different price structures. Flexible AC Transmission Systems (FACTS) devices also costs, relocatability etc. The impact of transmission congestions on electricity prices,
allow the increase of the overall utilization of an electrical power network by consumer, producer and society profit is also studied.
controlling the power flow.
The valuations carried out show that even if there is limited impact on removing
Since installations of FACTS devices require huge investments, with costs that could be congestions by installing a controllable device, it has a significant monetary value for
similar to new transmission lines, it is very important to investigate all different aspects the owner. In addition to quantitative results, qualitative aspects will describe the most
of such as a device: For example not only the increased transmission capacity, but also important factors when evaluating this value.

Partnerships: ETH Zürich

16 17

FLOWDEMO.NET: INTERACTIVE, INTERNET-BASED TOOL TO VISUALIZE POWER The user can interact with the calculation by switching on or off lines and components,
changing parameters such as the voltage of a regulating generator or the active and
FLOW IN AN ELECTRIC TRANSMISSION SYSTEM reactive power of a load. The GUI is programmed in Java.
Martin Geidl, Christian Schaffner
The calculation is carried out on a server that communicates with the java applet
Abstract running inside the Internet browser of the client computer. is an interactive load flow visualization software for education in
electrical engineering. What differentiates it from similar tools is that it can be run via Applications in Education
any Internet browser or, alternatively, as a stand-alone application. Thus there is no A typical application of is as following: The lecturer prepares
special installation needed for the user. Lecturers can use it to visually demonstrate demonstration networks and puts the according net lists on the server. During class he
load flow in electric transmission systems in a convenient way: They access predefined can explain theoretical principles using one of the prepared networks. All he needs is a
examples stored in a database on the server. Afterwards students can access the same computer with Internet access and a computer display projector. To strengthen the
examples via an Internet browser to deepen the new knowledge. knowledge students will use the same simulations after class to work through written
Description gives a visual representation of a calculated power flow including the New concepts or devices can be implemented directly by the lecturer, since the source
network topology, graphical representation of the load flowing through a line, code is available. Also, as Matlab is well known by most engineering people, no special
numerical and graphical representation of bus voltages and angles, state of switches programming knowledge is needed.
etc. The graphical user interface (GUI) of displays the network topology
together with graphical and/or numerical representations of the load flowing through New Version
a line, of bus voltages and angles, state of switches etc. During 2003 we brought online version 2.0 of which is a rewrite from
the ground up. The final version is available since later summer. It brings the following
improvements over version 1.0:
5.8MVar • A graphical network editor allows the user to design new networks or change
existing ones.
Bus 6
• The networks are stored in a central database, which makes it easier for lecturers to
publish new networks to students. This database has designated areas for each
Bus 5 -21.7MVar user (private and public) and for specific lectures.
• Flexible AC Transmission Systems (FACTS) such as SVC and TCSC can be added.
• A simple Market Simulator using optimal power dispatch makes it possible to study
the price-building mechanisms in a network. Cost curves can be assigned to each
Bus 4 generator and load and will calculate the amount and prices of the
38.29MW traded energy.
• There is now the possibility to store and load networks to/from the local disk
G -1.58MVar through the internet browser.
Bus 1
7.6MW • PDF-Printing makes it possible to add calculated networks into documentation
• A new compatibility mode for users behind a firewall

Bus 3
Success in education
Starting this winter semester was used successfully at different lectures:
-20.59MVar At KTH in Stockholm was used in an introductory lecture with more than
100 students, at ETH Zurich in the lecture "Elektrische Energiesysteme" to name a few.
Bus 2 has proven to be a stable system that is easy to use for educational uses.
Student evaluations are on-going.
Figure 1: A six bus network in Voltages, angles and load flows are displayed either in
numbers or graphically
Partnerships: ETH Zürich

18 19


SCHEME [1] Marek Zima: "Special Protection Schemes in Electric Power Systems – Literature
survey", Technical Report, Zurich 2002; available online:
Marek Zima
Deliberalization of the electricity market in many countries yields frequent changes of [2] Marek Zima, Göran Andersson, "Stability Assessment and Emergency Control
generation and transmission pattern. This together with the continuous grow of the Method Using Trajectory Sensitivities", presented at the 2003 IEEE Bologna
load demand causes significant stresses of the power systems, often beyond their PowerTech, June 23 - 26, Bologna, Italy, 2003.
limits. This in turn results in more frequent contingencies leading to severe [3] Marek Zima, Petr Korba and Göran Andersson, "Power Systems Voltage
consequencies – risk of failures, blackouts and financial losses. Emergency Control Approach Using Trajectory Sensitivities", presented at the IEEE
Conference on Control Application, 23 - 25 June, Istanbul, Turkey, 2003.
Since the extensions of the transmission systems are kept to a minimum for various [4] Marek Zima, Thilo Krause and Göran Andersson, "Evaluation of System Protection
reasons (mainly environmental), better utilization of existing assets becomes Schemes, Wide Area Monitoring and Control Systems", presented at the 6th
necessary. Various means can be used for this purpose, ranging from on-line APSCOM Conference, 11 - 14 November, Hong Kong, China, 2003.
monitoring of available transmission capacity to improved control, especialy under
Partnerships: ABB Switzerland
emergency conditions. The later one is task of System Protection Schemes (SPS) [1].
Conventional/traditional SPSs designs are based on the predefined worst scenarios
obtained by off-line simulations, which are used to define certain rules. These rules are
applied on tuning of local relays (underfrequency, undervoltage etc.) that shall operate
a way eliminating incipient instability. However, assumptions used for setting these
rules are too conservative and often leave a significant unused transmission margin.
Moreover, these schemes have more protection than control functionality and

Above mentioned disadvantages can be overcome by an on-line scheme, which would

take into account actual network conditions and provide high degree of coordination in
control. Such a SPS could rely on the phasor measurement units placed throughout the
network to be able to provide a wide area view, since majority of dangerous
phenomena in power systems have such an origin (voltage, frequency, small-signal

The goal of this research project is to determine constraints for physically feasible SPS
application, investigate and design an algorithm mitigating one or more instability

In the year 2003 the project activities were as follows.

• The new algorithm for emergency control has been proposed and presented in [2]
and [3]. The algorithm is a new concept introducing Model Predictive Control based
on the Trajectory Sensitivities.
• The long-term consequences of the Wide Area Control have been studied in terms
of the improved security of the power system.
• Assessment method for investment into a Wide Area Monitoring and Control
systems has been studied in [4].

This project is done partly at ETH and partly at one industry partner (which is actually
sponsoring the project) in order to take into consideration practical aspects of possible
future implementation.

20 21

3.3 Spin-Off Activities 4. Publications and Reports1

COMPANY PLEXIM 4.1 Publications
Since june 2002 the Power Systems Laboratory hosts the ETH spin-off Plexim. The
company was founded by Jost Allmeling and Wolfgang Hammer with the purpose of N. Dizdarevic, M. Majstrovic and G. Andersson
"FACTS-based Reactive Power Compensation of Wind Energy Conversion System",
bringing to market the simulation software PLECS that they had developed in the
Proceedings of 2003 IEEE Bologna PowerTech, Paper no 251, Bologna, Italy, 23 - 26 June
course of their research. After having developed PLECS into a marketable product the
company started selling the first licenses in November 2002. Today, the software has
M. Zima and G. Andersson
been sold to universities and industrial customers all over the world with a focus on
"Stability Assessment and Emergency Control Method Using Trajectory Sensitivities",
European countries and the United States. PLECS is intensively used for research and
Proceedings of 2003 IEEE Bologna PowerTech, Paper no 195, Bologna, Italy, 23 - 26 June
education activities in technical universities, especially in Germany and Switzerland.
Software PLECS Ch. Schaffner and G. Andersson
The initial reason for the development of PLECS was the lack of appropriate simulation "Determing the Value of Controllable Devices in a Liberalized Electricity Market: A New
tools for power electronic systems. Power electronic systems usually comprise both an Approach",
electrical circuit with many semiconductor switches and sophisticated controls. For this Proceedings of 2003 IEEE Bologna PowerTech, Paper no 276, Bologna, Italy, 23 - 26 June
type of system PLECS offers two major advantages compared with other simulation
M. Zima, G. Andersson and P. Korba
"Power Systems Voltage Emergency Control Approach Using Trajectory Sensitivities",
• PLECS embeds electrical circuits seamlessly into Simulink models. This facilitates 2003 IEEE Conference on Applications, CCA 2003, Istanbul, Turkey, 23 - 25 June
system simulations that comprise the electrical part modeled in PLECS and the
control structures modeled in Simulink. Thanks to the schematic editor of PLECS, J. Allmeling
entering electrical circuits is equally intuitive as modeling control structures with "Schnelle Kompensation von Harmonischen mit Aktivfiltern"
Simulink. Bulletin SEV/VSE July 2003, No. 15, pp.15-20
• PLECS uses ideal switches and accelerates the simulation of switched systems by a
factor of 10 to 100 compared with common circuit simulation programs. N. Dizdarevic, M. Majstrovic and G. Andersson
"Longer Term Stabilisation of Wind Power Plant Voltage/Reactive Power Fluctuations
The innovation of PLECS is the use of ideal switches. Unimportant detail information is by FACTS Solution"
neglected in order to obtain a high simulation speed. The simulation accuracy of the IASTED Power and Energy Conference, Marbella, Spain, September
entire system is not affected. A pleasant side effect is that the user is not confronted
with the specification of often unknown parameters. Therefore the initial effort to set M. Zima, Th. Krause and G. Andersson
up a simulation model is low and even inexperienced students will be able to work with "Evaluation of System Protection Schemes, Wide Area Monitoring and Control
PLECS. Systems",
Proceedings of APSCOM 2003, Hong Kong, China, 11 - 14 November
Future developments
PLECS is continuously improved by enhancing the component libraries and adding new A. Karpatchev, H. Glavitsch, and G. Andersson
features. In December 2003, Plexim has launched a KTI project in cooperation with ETH "Forced Symmetrization Techniques for the High Voltage Grid"
Zurich and ABB to expand the simulation algorithm in the next major release. (KTI = Proceedings of APSCOM 2003, Hong Kong, China, 11 - 14 November
Swiss Federal Commission for Technology and Innovation)
R. Sadikovic and G. Andersson
References: "Power Flow Control by Sensitivity Based FACTS Controllers"
[1] Proceedings of IPEC 2003, Singapore, 24 - 29 November

G. Andersson, T. Green, B. Pal and Ch. Rehtanz

"Advanced FACTS control"
ABB Review, 4 (2003), pp 21-26

Publications and Reports can be downloaded from

22 23

4.2 Reports 5. Presentations

G. Koeppel M. Zima
"Distributed Generation - Literature Review and Outline of the Swiss Situation", "Stability Assessment and Emergency Control Method Using Trajectory Sensitivities",
Technical Report, Zurich 2003; available online: presented at the 2003 IEEE Bologna PowerTech Bologna, Italy
June 23 - 26, 2003
Th. Krause
"Evaluation of Transmission Pricing Methods for Liberalized Markets - A Literature Ch. Schaffner
Survey", "Determing the Value of Controllable Devices in a Liberalized Electricity Market: A New
Technical Report, Zurich 2003; available online: Approach", presented at the 2003 IEEE Bologna PowerTech
Bologna, Italy
M. Milosevic June 23 - 26, 2003
"Decoupling Control of d and q Current Components in Three-Phase Voltage Source
Inverter", G. Andersson
Technical Report, Zurich 2003 "Regelenergie"
BKW-CEPE Workshop 03
Bern, Switzerland
4 July 2003

G. Andersson
European Energy Venture Fair 2003
Rüschlikon, Switzerland
28 October 2003

Th. Krause
"Evaluation of System Protection Schemes, Wide Area Monitoring and Control
presented at the 6th APSCOM Conference
Hong Kong, China
11-14 November 11 - 14, 2003

A. Karpatchev
"Forced Symmetrization Techniques for the High Voltage Grid"
presented at the 6th APSCOM Conference
Hong Kong, China
11 - 14 November, 2003

R. Sadikovic
"Power Flow Control by Sensitivity Based FACTS Controllers"
presented at IPEC 2003
24 - 29 November, 2003

24 25

6. Conferences and Visits

6.1 Conferences and Workshops

G. Andersson
Flexiges workshop, Comillas Unversity,
Madrid, Spain
14 January 2003

G. Andersson
Cigre B4-41 Meeting
Las Vegas
27 - 28 January 2003

G. Andersson
The Hong Kong Polytechnic University, Department of Electrical Engineering
Hung Hom, Kowloon
Hong Kong, China
17 - 21 February 2003

G. Andersson
ABB University
Power Technologies
Ludvika, Sweden
4 - 6 March 2003

G. Andersson
PHD Exam
Lund University, Sweden
24 -25 April 2003

G. Andersson
IEEE PES Fellow Committee
Minneapolis, USA
2 - 3 June 2003

G. Andersson, Ch. Schaffner, M. Zima

2003 IEEE Bologna PowerTech Conference
Faculty of Engineering , University of Bologna
Bologna, Italy
23 - 26 June 2003

G. Andersson
IEEE Power Engineering Society (PES)
General Meeting 2003
Toronto, Canada
13 - 17 July 2003

26 27

J. Allmeling, W. Hammer 6.2 Visits

EPE 2003 Conference and Exhibition
Toulouse, France
1 - 5 September 2003 Prof. Math Bollen
Chalmers , Sweden
G. Andersson 30 June 2003
38th Meeting and Colloquium of CIGRE Study Committee B4
Nuremberg, Germany Prof. S. T. Choi
21 - 26 September 2003 Kyongju University, South Korea
29 July 2003
G. Andersson, M. Zima
University of Liege Prof. J.-B. Lee
Liege, Belgium Wongkwang University, South Korea
1 - 3 October 2003 29 July 2003

G. Koeppel Prof. Ani M. Golé

DIgSILENT Workshop: Wind Power Integration Dept. of Electrical and Computer Engineering,
Gomaringen, Germany University of Manitoba
6 - 7 October 2003 18 - 19 September 2003

M. Geidl Rajiv Kumar

"Realität und Vision der ökologischen Stromversorgung" Uni. Karlsruhe
ETG/Electrosuisse, OGE/OVE, ETG/VDE 21 November 2003
Salzburg, Austria
BfE Meeting
5 - 6 November 2003
9 May 2003
A. Karpatchev, Th. Krause
IEEE Swiss Power Engineering Chapter
6th APSCOM Conference
12 June 2003
Hong Kong, China
11 - 14 November 2003 AGS Meeting
15 - 16 September 2003
R. Sadikovic
IPEC 2003 Optimization enabled Transient Simulation
Singapore, China 18 September 2003
24 - 29 November 2003

G. Andersson
CLP 6.3 Awards
Hong Kong
8 - 12 December 2003 M. Zima
Best Student Paper Award,
G. Andersson 2003 IEEE Bologna PowerTech Conference
PhD Exam Bologna, Italy
NTNU, Trondheim, Norway 23 - 26 June 2003
17 - 18 December 2003

28 29

Activities of the High Voltage Laboratory

1. Organisation
Head: Prof. Dr.-techn. Klaus Fröhlich

Secretary: Karin Sonderegger from 1 May 2003

Barbara Rutz until 31 May 2003

Scientific Staff: Dipl.-Ing. Stefan Berger

Dipl.-Ing. Andreas Bitschi from 1 May 2003
M.Sc.El.Eng. Mike Chapman
Dipl. El.-Ing. ETH Patrick Favre-Perrod from 1 May 2003
Dipl.-Ing. Manfred Grader
Dipl.-Ing. Martin Hinow from 1 May 2003
Dipl.-Ing. Wolfgang Hribernik
Dipl.-Ing. Bernd Klöckl
Dipl. El.-Ing. ETH Urs Krüsi
Dipl. El.-Ing. Evgeny Murtola from 1 May 2003
Dipl. El.-Ing. ETH Stefan Neuhold
Dipl. El.-Ing. ETH Diego Politano until 30 April 2003
Dott. Fis. Claudia Roero
Dr. sc. nat. Micha Semmler from 1 July 2003
Dipl. Phys. Ulrich Straumann
Dipl.-Ing. Ruben Vogelsang

Permanent Staff: El.-Ing. HTL Hans-Jürg Weber High Voltage Laboratory

Charles Sigrist Electronics Group
Heiko Vögeli Electronics Group
Henry Kienast Workshop

Scientific Associates: Tit.-Prof. em. Dr. sc. techn. Habibo Brechna

Dr. rer. nat. Timm H. Teich
Prof. em. Dr. Ing. Walter Zaengl

Visiting Lecturers: Dr. Werner Hofbauer, ABB High Voltage Techn. Ltd.

30 31

2. Teaching
The lectures and laboratory classes listed in the following section are part of the
standard curriculum of the Electrical Engineering Department and are conducted by
the staff of the High Voltage Laboratory. Details of the entire electrical engineering
curriculum can be provided on application (list of lectures, option proposals).

2.1 Lectures

1st Semester 2V+2U

Networks and Circuits I Fröhlich K.
Netzwerke und Schaltungen
The electric current and voltage; linear and non-linear resistive circuit elements; theory
of meshed linear circuits (time variant and invariant); electric energy and power; ideal
amplifier circuits with controlled current sources. Non-linear resistive systems,
transistor amplifier as a non-linear system.
As in the first semester the mathematical basics are not yet fully developed, the lecture
is limited to direct current circuits. The knowledge to be achieved will be intensified by
through exercises.

5th semester 4G
Electric Power Systems Andersson, G.
Elektrische Energiesysteme Fröhlich, K.
Structure of electric power systems; symmetric three phase systems; modelling of
power transformers and generators; analysis of plain and unsymmetrical three phase
systems; transient switching phenomena; basics of current interruption; principles and
applications of distribution- and transmission switchgear; basics of insulation

6th or 8th Semester 4G

High Voltage Technology Fröhlich, K.
Basic phenomena related to gaseous, fluid and solid dielectrics; dielectric breakdown
mechanisms; dimensioning of high voltage components by employment of theoretical
considerations and computer modelling tools (small project); sources of overvoltages
(switching and lightning); overvoltage protection; investigation of dielectric stresses by
computer modelling (small project); insulation co-ordination.

7th Semester 4G
Technology of Electrical Power Systems Fröhlich, K.
Technologie elektrischer Energiesysteme
Basic physical aspects when conducting high current and high voltage for distribution
and transmission of electrical power; emerging technology in distribution and

32 33

transmission systems (super conductivity, fault current limitation, energy storage, 2.2 Student Projects
HVDC); electromagnetic compatibility for system and personnel; intelligence of power
system equipment (control, model based diagnosis); decentralised, renewable energy
sources; project work; excursion to a utility and to a manufacturer. To be admitted to the diploma examinations of the 7th and 8th semester, students of
the electrical engineering department have to carry out two projects. Each student can
freely choose his subject area, but usually the two projects have to originate from
7th Semester 2V different subject areas. According to the curriculum, two days of the week during the
Engineers Work – Technique and Economics Hofbauer, W. semester period are to be devoted to this work. In general, the subjects are derived
Ingenieurarbeit - Technik und Wirtschaft from topical research and development tasks.
After a short introduction to purpose of an enterprise, its control and the role of the
engineer will be explained by the example of surge arresters. By some examples the M. Bartholet Aufbau und Test eines Transformatorprüfstandes zur
accounting principles will be presented focusing on the meaning and goal of the Ch. Zwyssig experimentellen Identifikation von Prozessmodellen für ein
financial statement, the income statement and the balance sheet. The importance of Diagnosesystem
the capital expenditure accounting is explained which considers besides product Th. Brügger Ausbreitung elektrischer Entladungen an geschädigten Glimmer-
related cost factors like functionality, design and variety of variants, also process bändern
related cost factors like personnel, infrastructure and make or buy decisions. By specific
V. Bürker Rechnergestütze Berechnung kritischer Designparameter zur
consideration of the engineers’ work the importance of the Research and Development
Entwicklung von Löschkammern für Hochspannungsleistungs-
process and its impact on the success of an enterprise will be explained.

5th - 8th Semester 4G E. Gallone Water droplets on High Voltage Conductors under a 20kw/a
Computer Science Oriented Project Work Fröhlich, K. Electric Field
EDV-orientierte Projektarbeit and assistants S. Guha Thakurta Calibration and modelling of oil humidity sensors for online-
monitoring of power transformers
Using information technology tools, the students, operating in teams and with only
limited supervision, have to find solutions to topical problems chosen from the M. Nibelle Demonstration of the suitability of electromagnetic emissions for
research or teaching activities of the high voltage group. the measurements of aring time in circuit breakers
Depending on the tasks, existing programme packages may be applied or, if necessary, Ph. Simka Dimensionierung und Aufbau einer Hochstromversorgung zur
new programmes or programme subsections have to be compiled Lichtbogenerzeugung

2.3 Diploma Projects

Allocated time is four months. The majority of students devote their time to this work
in the winter semester. The student has the option to carry it out either before or after
the formal diploma examination (dates in spring and autumn).

M. Lombardi Bestimmung der Restfestigkeit sowie der Isolationsstruktur an

L. Pedrini betriebs- und zeitgealterten Wicklungsstäben von rotierenden
(EIF Fribourg) elektrischen Maschinen
M. Wagner Determination of Life-Cycle Cost Savings due to On-Line
M. Flubacher Monitoring of Circuit Breakers
(Diploma Project in Brisbane/Australia)
L. Walter Resonance characteristics of a 100MJ/100MW SMES coil
experimental model verification

34 35

2.4 Excursions 3. Research Activities

ATEL Olten, 20 January 2003 3.1 Completed PhD Theses
Visit of net control center (Netzleitzentrale) and trading floor

Nordostschweizerische Kraftwerke (NOK) in Thalwil, 29 January 2003 THE PARTIAL DISCHARGE BEHAVIOUR IN MICROVOIDS OF POLYMERIC
Visit of a 220kV GIS (Gasisolierte Schaltanlage) INSULATION MATERIALS
Visit of a GIS (generator circuit breaker)
Candidate: H.-P. Burgener
Energy Excursion 2003, 20 June – 22 June 2003 Thesis: ETH No. 15028
Visit of Nexans Suisse, Cable Fabrication, Cortaillod, Switzerland Date of oral examination: 26 Februar 2003
Visit of Power Station Grimsel, Grimsel, Switzerland Examiner: Prof. Dr. K. Fröhlich, ETH Zürich
Co-examiner: Prof. Dr. K. Feser, Universität Stuttgart
ABB Hochspannungstechnik, Baden, Switzerland, 23 June 2003
Visit of a MO, surge arrester (Ueberspannungsableiter) Author’s Summary
Fibre-reinforced materials are widely used in the field of power engineering. Due to
ABB Hochleistungslabor, Baden, Switzerland 4 December 2003 their outstanding mechanical and dielectric behaviour these materials find also high
voltage applications. The dielectric properties of the fibre-reinforced insulation
materials are strongly related to the manufacturing process. The continuous
development of the process engineering suppresses the bigger gas-filled voids of the
order of one millimetre and the related partial discharges (PD). Improved material
properties lead automatically to designs employing increased stresses. Thus the PD
behaviour of weak spots in the order of micrometres such as delaminations between
fibre and matrix or cracks can be of importance in the future. However, such small
voids could lead to very small partial discharge levels (< 1 pC), which could be hardly
detectable under practical conditions due to ambient noise, but could nevertheless
affect the insulation.
The reasons stated above demand a closer investigation of the PD behaviour of
microvoids. The literature available shows that so far a range of experiments and
simulations have been accomplished to assess the PD behaviour as a function of the
geometry of disc-shaped or spherical voids. However, the influence of the side walls on
the discharges has not been taken into account within these investigations. In contrast,
this work considers the restricting effect of the side walls on the discharge with narrow
microvoids or cracks parallel to the electrical field. Such cavities can also appear within
fibre-reinforced materials. Considering this aspect methods were established, which
distinguish themselves by an improved sensitivity and which could support the
conventional PD measurement in assessment or dimensioning of insulation materials
in the future.
The restricting influence of the side walls of longish voids parallel to the electrical field
has been investigated by means of an experiment. To assess this, well-defined holes of
constant length (300 µm) and various diameters (50 – 300 µm) were drilled into the
epoxy resin of a resin-filled electrode system. For each diameter the partial discharges
were measured electrically as well as optically. A considerable increase of the PD
inception voltage and inception field strength was found towards smaller diameters
(< 100 µm). This increase has its origin mainly in the diffusion-based interaction of the
electrons with the side walls.

36 37

The experimental effort for the manufacturing of the samples was large and the STRESS DEPENDENT ECONOMIC ASSESSMENT OF INNOVATIONS IN POWER
investigations of the PD behaviour of even narrower voids (< 50 µm) of various lengths
were restricted due to the methods available for the preparation of the defined holes.
Considering that voids of smaller dimensions can not yet be assessed by experiment a Candidate: Diego Politano
model for the simulation of the PD inception in narrow, cylindrical voids has been Thesis ETH No. 15183
provided. The model is based on the continuity equations, which were coupled via the Date of oral examination: 27 June 2003
charge densities in the discharge with the Poisson equation and the electrical field. The Examiner: Prof. Dr. K. Fröhlich, ETH Zürich
PD inception criterion is based on a photonic feedback, i.e. a Townsend mechanism. Co-examiner: Prof. Dr. D. Birtwhistle, Queensland University,
With the experimental fitting of a parameter, which is describing the interaction of the Brisbane, Australia
electron avalanche with the periphery of the cylindrical void, a good correlation
between the calculated and the measured PD inception voltages could be obtained. Author’s Summary
For the application of the model presented as a tool for an electrical dimensioning, the When a power utility considers the introduction of a new component in their
materials-specific geometries of the voids must be known and modelling with a substation, it is primarily faced with costs. Each proposed investment must be
cylindrical geometry must be applicable. Thus the usefulness of the simulation was defended facing the management. In the simplest situation a new component with
investigated with mechanically stressed polyester fabric reinforced epoxy tubes. The higher performance will just be substituted for an older one. In this case the existing
light microscopy served to assess the mechanically induced geometries of the cracks. By methodologies are sufficient to assess the economic impact of the investment. In many
means of comparison of the calculated and optically measured lengths of the cracks cases a simple comparison is not possible because the new substituted component has
the applicability of the simulation of the PD inception field strength could be shown. very different properties or because the innovation is simply added to the substation,
For a simulation of the apparent charge the existing model should be expanded. Thus a thereby increasing substation complexity. For these cases the costs caused by
model presented in the literature [1] has been used. innovation must be justified by the overall benefits brought to the substation. A
Optical investigations of voids specific to the material in combination with simulation methodology for the evaluation and the calculation of the financial impact of a
tools offer the potential to consider critical field strengths as well as void geometries, in modification in power substations is presented in this work.
a range not accessible to the conventional PD measurement. Thus the methods
presented could be very helpful to users and manufacturers of insulation materials for The new idea compared to the conventional cost calculation methodologies is that the
formulating their requirements/specifications. life-cycle costs of power apparatus are evaluated in dependence on the stresses
Further investigations showed the influence of the sequential as well as simultaneous effecting an apparatus. The overall economic impact of an innovation is calculated by
mechanical and electrical stresses on glass fibre-reinforced epoxy tubes. The investigating its implications on the various items of apparatus of the substation.
experiments revealed that the dielectric failure potential is higher with simultaneous When subjected to certain stress conditions during a period of time, the new apparatus
mechanical stresses. The reason is the generation of charge carriers during the cracking may suffer slight damage. The consequence is that an increased probability of failure
of the barrier layer between fibre and matrix. These charge carriers combined with a compared to the initial state is possible. Subjected again to new stresses, the apparatus
high field strength could lead to an avalanche development and to the dielectric failure. condition gets even worse, with an increased risk of failure. If no maintenance is
In the mechanical dimensioning and specification of fibre-reinforced materials, which performed this process leads sooner or later to the end-of-life of the apparatus. A
have to support high dielectric and mechanical stresses, it has thus to be taken into mathematical model for the description of such a life cycle is described in this work.
account that even small-scale delamination and cracking must be avoided. The model is based on the identification of all possible condition-states of an
apparatus, the calculation of the probability of the change of condition and the
Partnerships: ABB High Voltage Technologies Ltd, Zürich, Switzerland association of all of these condition-states to cost parameters. The analysis is
Cellpack Ldt., Wohlen, Switzerland subdivided into seven phases. First the apparatus is broken down into its vital
Micafil Ltd., Zürich, Switzerland
subcomponents in order to identify all the condition-states leading to a failure in the
Reference apparatus assessed. This information gathering is carried out performing a Failure
Modes and Effect Analysis (FMEA) and a Fault Tree Analysis (FTA) on the apparatus. In a
[1] G. C. Crichton, P. W. Karlsson and A. Pedersen, "Partial Discharges in Ellipsoidal and next phase the designed life-cycle process of the apparatus represent a multistate
Shperoidal Voids", IEEE Trans. on Electrical Insulation, Vol. 24, No. 2, Apr. 1989, pp. system and is modelled with a Markov chain. The stresses acting on the apparatus are
335-342. investigated and the correlated state-transition probabilities are inserted. The Markov
process is solved. To define the costs related to the condition states of the apparatus a
similar algorithm is applied. Every condition-state of a subcomponent is associated
with costs. Finally, the cash flow stream related to the investigated investment is
calculated and the final decision about the profitability of the investment can be taken
on the basis of the LCC calculation made.

38 39

To prove the suitability of the algorithm, the methodology has been applied in the 3.2 Current projects
assessment of two case studies. In the first one the economics of controlled
energization of power transformers is investigated. The second one deals with the
financial impact of the monitoring and diagnosis of HVAC circuit breakers. The step-by- GECKO - GEOTHERMAL ENERGY CONVERSION KEY OBJECTIVIES
step application of the algorithm has improved the understanding of the technical and
economic impact of both innovative technologies. The models for the life-cycle costs of Andreas Bitschi
a power transformer and a circuit breaker have been developed. In both case studies Goal of the project
the methodology helped to give evidence of a financial benefit due to the introduction While the emphasis of geothermal research is on the rock characteristics and the
of the innovations. It has been concluded that: development of the reservoirs, other problems under study are concerned with the
• The uncertainty of many of the input parameters is a major weakness of the stress- practical use of geothermal heat. Apart from the well-known use in geothermal central
dependent life-cycle costs methodology. In particular, the condition-state- heating installations the transformation of the warmth into electricity is a priority goal.
transition probability can often only be estimated on the basis of limited historical In order to effect this different processes and plants are available.
data. Therefore it is necessary to create a worldwide database containing All those considered here comprise a turbine, which drives via a shaft a generator. The
information about the condition state transition probabilities, the cost parameters fluid issuing from the borehole serves either directly as working medium or it transfers
and failure mode analysis of power components. the warmth in a heat exchanger to a secondary fluid (binary systems).
• In order to reduce the uncertainty of the input parameters the methodology should As the geothermal energy from thermodynamic view represents low-temperature
be primarily used in comparative studies. (~200°C) energy and thus contains a large portion of energy, which cannot be
• Sensitivity analysis and risk assessment procedures are helpful in dealing with the converted into technically useful energy, the Carnot-efficiency of the transformation is
uncertainty related to the input parameters. In addition performing sensitivity relatively limited. At this time Organic-Rankine-Cycles (ORC) and Kalina processes have
analysis the assessment of the weight of each parameter over the total costs leads the highest efficiencies.
to the identification of those parameters on which it is worthwhile investing time In the context of the project methods are to be investigated for the transformation of
and money to increase the exactness of their values. low-temperature warmth into electricity with relatively high efficiencies. For this in a
first step the state of the art was checked, now we are looking for improvements and
Future development should focus on the investigation of the life-cycle processes of proceed with the development of new ideas. A next step is the combination of the
other apparatus in substations and on the application of the methodology for the different processes and the identification of synergies. The last step would be the
investigation of different substation configurations. integration in existing and new power systems.
A further aspect of the task is the downhole principle. That means the energy
transformation should happen in the borehole. Therefore a machine without any
moving parts would be very advantageous because the chances of maintenance will be
From thermoelectric to thermionic
Heat exchangers incorporating thermoelectric modules offer one possibility to
generate electrical power from geothermal sources. These devices convert a
temperature difference into an electric current. (Seebeck-Effekt). Thermoelectric
materials are expensive and there is not more than 5-8 % efficiency available under
ideal conditions. This circumstance is caused by the thermal conduction through the
semiconductor crystals.
An application that reduces the heat transfer to thermal radiation is the thermionic
converter. Here a vacuum gap between the electrodes eliminates the heat conduction
and the current is generated by thermal emission of electrons. For these electrons the
vacuum gap means a potential barrier, which they have to surmont. The required
energy of the electrons would need temperatures > 1000 K.
An additional problem ist the fact of space charge, which further increases the height
of the potential barrier.

According to the quantum mechanics it is also possible for the electrons to tunnel the
barrier depending on the height and width of the barrier. For this a barrier thickness of

40 41

nm-scale is required. For the implementation of such small gaps, special fabrication conductor. This energy carrier medium is heated up by the ohmic losses in the electrical
methods and positioning systems are needed. A further way to influence the shape of conductor. The layout of this bus offers several degrees of freedom: among them the
the energy barrier is an electric field between the hot and the cold sides, respectively, of type of coolant (gaseous, liquid), the diameter of the inner conduit and the operating
the electrodes. In this case the superpositioning of the potentials leads besides a pressures.
decrease of the barrier-height also to a reduction of the width of the barrier, so that the
electrons already can tunnel on the height of the Fermi level.

For tunneling the required energy of the electrons is much lower than in the case of
overcoming the barrier and so can be effected by a lower temperature source like a
geothermal reservoir.
Outlook 2004
• Research with partners in materials science
• Analytic calculations and modelling Electrical energy
Chemical energy
• Numerical simulations of static converters
Heat flux

Partnerships: under negotiation


Patrick Favre-Perrod
The project replaces the former FUEL project and is a part of the broader Vision 2020
project. The project aim is to formulate proposals on how to implement energy Figure 1: Future energy networks will connect traditional plants as well as new distributed producers to
consumers. The energy exchange is likely to be electrical, chemical and thermal.
transmission in the envisioned future energy networks
As a first result, it has been established that liquid coolants
According to our analysis, an evolution in transmission technologies will be induced by
• require lower operating pressures
several constraints:
• allow for higher power densities
• Due to their impact on the landscape and EMC issues, as well as the limited
available rights of way, new overhead lines are unlikely to be built. • offer a cooling power which is proportional to the temperature rise of the coolant
• The economical operation of distributed generators will need solutions for (low)
whereas gaseous coolants
power infeed into the transmission network. The implications on the grid hardware
might be lower operating voltages, d.c. operation, etc. • enable a decoupling of the transmitted chemical power and the absorbed waste
• Future producers and consumers (microturbines, fuel cells, etc.) will not handle
exclusively electrical energy. Technological solutions using synergies between • can absorb much loss energy at moderate outlet temperatures
different energy forms (heat, chemical energy, electricity, etc.) will eventually be • enable a limited storage of chemical energy in the transmission infrastructure.

A possible structure of the envisioned energy network is shown in Figure 2. Different

consumers and producers might connect to "energy hubs", which will permit energy
exchanges as well as storage of excess energy, e.g. in chemical form. It appears that
this grid provides more flexibility if these "hubs" can exchange these energy forms over
longer distances. To profit from possible synergies, a combined energy transmission
"bus" for several energy forms might be useful.
Figure 2: Principle layout of hybrid energy transmission device
As a first theoretical approach an energy bus for thermal and electrical energy has been
studied (Figure 2). A chemical energy carrier flows through a coaxial electrical Partnerships: ABB Switzerland, Alstom France, ETH Zürich

42 43

STORAGE IN STOCHASTICALLY FED ENERGY SYSTEMS These two steps make it possible to calculate the power demand signal a storage
device would face, analyse it and feed it into the physical models of single storage
Bernd Klöckl devices or cascaded configurations. Proceeding in this way, one can assess the exact
storage demand for a certain situation, not only from the energetic point of view, but
Aim also from the peak power point of view. Figure 1 and 2 show the systematics.
To find technology-dependent optimum energy storage sizing algorithms for future
systems. 500

Linked to the "Vision 2020 (Vision of Future Energy Networks)"-project, this research 400

work covers the technological storage aspects in stochastically fed energy systems. day night
Energy storage integration is the obvious procedure of choice to overcome the inherent 300
problems raised by stochastic sources in electric power systems. However, location as

power [kW]
well as optimum sizing and technological implementation of storage devices has
always lacked a systematic theoretical coverage. 200

Recent publications about energy storage have proposed a new view on the problem by
developing a correct (although simplified) dynamical model of different devices into 100
normalised correlations between power density and energy density, which are the two
most important parameters to assess storage technologies [1,2]. This simple and
efficient idea helps to understand the applicability of a technology for a certain 0

constant power application. This approach is being extended to a more general one,
where power demand is a function of time. This situation applies particularly in energy −100
380 390 400 410 420 430 440 450
nodes (hubs of electric and other forms of energy with power infeed, consumption and time [h]
integrated storage capability) which are fed by stochastic sources. It has been found
out that, not surprisingly, the achievable energy density of a storage device exhibits Figure 2: The power curve of a large photovoltaic installation during three winter days1. The shaded areas
represent energy contents. The envelope curve is shifted vertically by the yearly average
remarkable variation with different power demand functions. Currently, it is power of the system which has been computed by a trend analysis algorithm.
investigated what kind of functions the power signals would impose on a storage
device located close to a stochastic source. The investigation procedure must therefore Figure 2 shows the contradicting requirements for storage technologies in
include the collection and correct interpretation of P(t)-functions of many different stochastically fed energy systems. While during daytimes the storage is required to
kinds of sources (note that in this generalised approach a stochastic sink is viewed as a absorb very dynamic high power signals from the PV arrays, it has to re-inject almost
negative stochastic source) and the development of average power calculation constant power into the grid during night time if it is designed to produce constant
procedures. baseload for an arbitrary system. Both of these exigencies can generally not be covered
by only one storage technology since the general basic distinction can be made into
Input: time-domain stochastic data of source or sink only two principles: Potential and kinetic storage. The determination of the exact
optimum physical implementation by analysis of such signals is the basis of this

probability analysis t/f – Analysis
information on t lost!
Outlook for 2004
Output: Time-frequency analysis of stochastic sources will be integrated into energy storage
+ Storage + theory. Combinations of storage devices will be investigated with various power
demand curves and optimum cascading algorithms will be developed. The result

should imply requirements for future energy storage technologies, e.g. high energy
Wdel @ P = const Wdel @ P = P(t)
„Ragone plots“
Transmission transmission cables or optimised energy nodes. The correlation with future
transmission technologies (see also figure 1) and the investigation of possible synergies
conventional HVL approach will be in the focus and form the interlink to P. Favre-Perrod’s scientific work at the
High Voltage Laboratory.

Figure 1: The HVL approach for the determination of optima in storage technology
Data for these investigations was kindly provided by HTA Burgdorf. Internet:

44 45

Further activities substantial damage. The investigation therefore centres on how the controlled
In coordination with the work of A. Bitschi, a publication about a transverse flux linear switching technology avoids inrush currents and reduces the stress profile for the
generator design for geothermal power conversion was elaborated ([3]). Additionally, B. transformer. It is fair to ask the question in which application fields the additional
Klöckl extended his collaboration with the University of Leoben, which resulted in investment cost for the switching technology is balanced by the extended lifetime of
another publication about permanent magnet synchronous machines ([4]). During July the transformer.
2003, the High Voltage Laboratory hosted the academic guest Prof. Jong-Beom Lee
from Wonkwang University, South Korea, who was attended by B. Klöckl. As a result of Maintenance strategies
this visit, a publication in the field of time-frequency analysis of transients in power Power companies want to reduce their maintenance cost by eliminating any
systems is to be submitted to IEEE (preliminary working title see publication [5]). superfluous maintenance schedule. However, a lower maintenance frequency can lead
to a higher failure rate of the components. In this case the penalty cost for the energy
References not delivered and higher investment cost for the earlier replacement of the component
are to be taken into account.
[1] T. Christen, M. W. Carlen: Theory of Ragone plots; Journal of Power Sources 91 The questions to be asked is: what is the cost with the present day maintenance
(2000), 210-216 schedule and what is it with an optimized maintenance strategy and in which
[2] T. Christen, C. Ohler; Optimizing energy storage devices using Ragone plots; application fields it is useful to change the strategy to the condition-based
Journal of Power Sources 110 (2002), 107-116 maintenance?
[3] B. Klöckl, K. Fröhlich: Energiekonversion Downhole – eine Vision; presented at the
Electrosuisse PES meeting "Geothermal Energy Conversion", Brugg, Switzerland, Substation layout modification
19/3/2003 One very well known controversial subject is how much redundancy should there be in
[4] A. Thaler, B. Klöckl: Measurement results using the fictitious field current model the optimized substation layout. Highly redundant structures have high investment
for PMSM; presented at the European Power Electronics Conference, Toulouse, and maintenance cost. The additional components (double bus bar in a GIS for
France, 2-4/9/2003 example) have to be maintained. On the other hand, the higher availability reduces the
[5] K.-H. Kim, J.-B. Lee, B. Klöckl, K. Fröhlich: Wavelet and neuro-fuzzy based fault possibility of incurring high penalties.
location for combined overhead/cable transmission lines (working title); to be
submitted to IEEE Trans. Power Delivery State of the art of the project
A methodology to calculate the stress dependent life cycle cost for one component has
Partnerships: ABB Switzerland, Alstom France, ETH Zürich been developed. It includes a Markov process which comprises four condition states.
This model has been applied in the assessment of two case studies. One deals with the
economic benefit of controlled energization of the power transformer, the other deals
with the impact of the monitoring and diagnosis system of the circuit breaker. The
ECONOMIC ADVANTAGES OF INNOVATION IN POWER SYSTEMS result of both case studies was that the innovations were amortized within a few years.
Martin Hinow Outlook for 2004
The next step is to enhance the model to handle configurations with more
Goal of the project components. Therefore the methodology has to be combined with a sensitivity
The liberalization on the energy market forced the energy suppliers to optimize the analysis. Furthermore there is a need to develop methodologies to model aging
costs in their system. The substation is one of the most cost sensitive points in a power processes and different maintenance strategies. One possible methodology may be
network. The investigation of the project focuses therefore on the Life Cycle Cost of a found in the determination of maintenance-dependent reliability characteristics.
substation. The cost evaluation deals not only with the components to be considered.
Usually a modification in the substation layout or a replacement of a component has Partnerships: ABB Switzerland, Alstom France, NGC England, RWE Germany
consequences for the overall expense. A simple comparison of cost is thus not possible.
The substation has to be considered as a complex system. The goal of the project is to
develop a methodology for LCC-calculation, which takes the substation complexity into
Stress dependent LCC
The power transformer is one of the most important and most cost intensive items of
equipment for electrical power transmission and distribution. If a fault in a transformer
occurs, this causes not only an interruption of electrical power but also large economic
losses. From failure experience on transformers it is widely suspected that inrush
currents, occurring when energizing unloaded transformers, was the reason for

46 47

MODEL-BASED TRANSFORMER DIAGNOSIS An experimental setup consisting of a 630 kVA distribution transformer equipped with
a controllable loading setup provides the functionality for heat runs (see Fig. 2). Sensors
Wolfgang Hribernik for temperatures, load and oil humidity deliver signals that are recorded with a data
logger. The setup is used to adapt the model with model verification procedures and
Goal of the project furthermore to develop procedures for implementation of the diagnosis system at a
Power transformers are one of the most expensive investments in an electric power power transformer in the field.
system. The deregulation of the electricity market forces the power utilities to reduce
maintenance costs. This leads to a lack of transformer specific knowledge inside a

oil temperature (°C)

power utility and increases the risk of a transformer failure. Moreover, the changed 80
market situation increases the demand for controlled transformer life-time 60
In order to estimate the actual condition of a power transformer, diagnostic
measurement techniques such as dielectric spectroscopy, frequency response analysis, 20
partial discharge measurement, etc. are in use. For the interpretation of the results and 0
based on that a decision what further action is called for requires knowledge about 0 20 40 60 80
relations between the measured properties – a model. Time (hours)

oil humidity rel. (%)

Sensors e.g. for transformer temperatures, oil humidity, gas in oil, etc. provide data
reflecting the actual transformer condition online. However, intelligent procedures for
data interpretation are needed in order to gain information about the transformer 5
condition from the data sets delivered by the sensors.
Combining models for offline and online measurement techniques provides synergies,
either for the modelling process or for decision making. 0
0 20 40 60 80
time (hours)
Model structure for a model-based diagnosis system
For the thermal and dielectric behaviour of a power transformer, a model structure Figure 2: Measured values of transformer oil temperature and relative oil humidity during a heat run
providing diagnosis facilities has been developed (Fig. 1). The thermal subsystem
calculates transformer temperatures from load and ambient conditions. A dielectric
Outlook for 2004
subsystem predicts the (nonlinear) diffusion process of water from transformer oil to
transformer paper. Dielectric spectroscopy provides information about the initial • Improved modelling of the diffusion process of water between transformer oil and
moisture content of the transformer. transformer pressboard.
The predicted values are compared with data obtained online by means of sensors. This • Development of adaptive strategies for identification of the temperature
redundancy gives the opportunity to distinguish between system changes, sensor dependent diffusion coefficient.
faults and modelling errors. • Modelling of the static and dynamic behaviour of capacitive-type sensors for
relative oil humidity.
• Validation of nonlinear models and self-learning systems using the subsystems of
ambient temperature Ta To the model-based diagnosis system.
cooler status c thermal
load S Partnerships: PSEL
To EPF Lausanne
residual generator

Chalmers University of Technology Gothenburg, Sweden

dielectric ho decision
paper moisture hp BPA Bonneville Power Administration, Portland OR, USA
subsystem hp
oil temperature To
diel. response function f(t)
paper temperature Tp dielectric hp
oil temperature To subsystem
oil conductivity γo (offline)

Figure 1: Model structure for a model-based transformer diagnosis system. A residual generator
compares measured and predicted values. Redundancy between some residuals provides the chance to
distinguish between transformer faults, sensor faults and modelling errors.

48 49


Manfred Grader, Stefan Berger 1250
1000 Ir = 2500A
Goal of the project
The fundamental design of today’s high-voltage circuit breakers has not changed for HV
Ur = 12kV Ur = 72.5kV
several decades, and it appears that the limits of improvement of this technology have 625
almost been reached. The main points of the NST project are the concept and design of Ir = 1250A

new, unconventional switching principles and the exploration of the physical and
MV Ir = 800A
technical fundamentals of their elements. Particular attention is paid to the
requirements of potential configurations of power networks of the future.
Partnerships: ABB Switzerland
300 1800 3600 4000
6000 n
Figure 1: Number of contacts connected in series (n) and the number of parallel paths (m) for a current-

Manfred Grader Although the number of contacts is pretty high, the space required is not the problem.
But considering that the costs for one elementary switch are in the range of 1 CHF,
Introduction building up a matrix of such dimensions is not practicable.
Based on experiments with a small number of low current contacts connected in series
and parallel, calculations have been made to estimate the requirement of numbers of
contacts to implement circuit breakers with a high switching capacity. Thought has number of contacts (x1000) volume (m3)
been given to the required space, the power losses and to the costs.
Number of contacts for successful switching
The basic values of the contacts in the calculations are taken from the properties of current-zero breaker 83.2 - 570 1400 - 2930 0.12 m3 0.63 m3
relays [1] we have used for building up the matrix in the experimental setup. The
parameters of the rated voltage, rated current and the short-circuit current are taken current-limiting breaker 120 - 1125 2250 - 7625 3
0.24 m 1.6 m3
from standardized values of the medium voltage (MV) and high voltage (HV) domain.
Figure 1 shows the required number of contacts connected in series and the number of
Figure 2: Total number of contacts and dimensions of the matrices
parallel paths to build up a current-limiting circuit breaker.
The number of required parallel paths m is given by the current, carried by the matrix. Also an analysis of power losses has been made. The calculations point to the fact that
Given that the short-circuit current is limited by the breaker, the values of the rated the losses are linearly dependent on the number of contacts connected in series.
current Ir result in the number of parallel paths in the MV and HV regime. The number Furthermore, the losses of a current-limiting breaker are pretty high (some hundreds of
of contacts n is defined by the fact that the short-circuit current is limited by the sum of kW), due to the fact that we have a low number of parallel paths and a high number of
the arc voltages (20 V) of the opening contacts connected in series. contacts connected in series.
Calculations have also been made for a current-zero circuit breaker due to the fact that
the boundary conditions are completely different. In this case the number of parallel Outlook for 2004
paths is given by the ability to carry the short-circuit current for one half-wave and the • Next step is now to define the necessary standards of an elementary switching
number of contacts connected in series is influenced by the required withstand voltage system for creating a practicable matrix out of these calculations.
of the breaker and the electrical strength of the gap of an open contact. • Estimation of the possibility to implement a practicable matrix with other
Figure 2 lists the total number of contacts for a matrix implemented as a current- technologies (e.g. nanotechnology).
limiting circuit breaker and a current-zero circuit breaker. The dimensions of one
elementary switching system (contact, drive and condenser) leads to the space required References
for the matrix.
[1] AXICOM data sheets, FX2 Relay,

50 51

FAST SWITCHES AND HYBRID COMBINATIONS lower even at relatively high circuit impedances in the commutation path. At contact
velocities greater than 20 m/s the arc voltage increases almost linearly with time,
Stefan Berger which is in good agreement to a blackbox model after Ayrton.
Research on fast switches and their behaviour in hybrid switching systems has been
followed up. Up to now adequate knowledge of the arc physics and electrical behaviour
for contact separation speeds grater than 10 m/s seems to be absent from the
literature. Measurements with fast switches in various circuit configurations were
therefore performed, especially focused on the current commutation in hybrid systems.
Besides the electrical characteristics of the arc during commutation, it is of particular
interest, whether there exists a limit for arcless commutation and if so one wants to
know its dependence on the circuit parameters and the contact separation speed. A
new fast switch has been developed to reach higher velocities (50 m/s) of the moving
contact. This switch has one breaking contact and a low impedance in the current path.
It has been placed in a test arrangement for alternating currents up to 10 kA at voltages
in the range of 50-1000 V.
Test arrangement
A synthetic test circuit has been built to measure the current commutation behaviour
at different contact speeds with variable parameters of the commutation path (Fig. 1)
The total current has been measured with a current transformer, the current in the
parallel path with a 1 m shunt. A differential probe has been used for the voltage
measurement. The measurement arrangement has also been improved due to
potential separation and lower current coupling resulting from the drive coil of the fast

i_tot switched parallel

Lt Rt path (fs) path (par)
Figure 2: Typical waveforms during in the commutation interval. Parameters: v = 23 m/s, Lt = 230 µH,
Ct = 43 mF, Rt = 32 mOhm; (a) Lpar = 3.2 µH, Rpar = 72 mOhm; ( b) Lpar = 1.7 µH, Rpar = 49 mOhm
Lfs Lpar
By now the limit for arcless current commutation could not be detected. Theoretical
u_fs considerations indicate that this limit is strongly dependent on the commutation
Ct Rfs V Rpar circuit impedance rather than the opening speed. None the less the fast switch offers a
commutation time of less than 500 µs.
Fast Switch Outlook for 2004
Rshunt • Complete the measurements and data analysis of the commutation process
• Investigate the influence of the contact material
• Derive a physical model that suits the measured data as well as known theory
Figure 1: Schematic of the test circuit • Modelling of the system behaviour of rapidly elongated arcs
Figure 2 shows typical waveforms during the commutation process for two different
circuit configurations. One can identify a phase of sliding contacts between 90 µs and
590 µs after the beginning of the current flow in the drive coil as a result of the contact
geometry. In this phase the arc is very short and the voltage drop is determined by the
electrode falls. Thereafter an arc burns as optically detected, which is confirmed in the
voltage and current waveforms. The commutation time is about one millisecond or

52 53

MATIC – MEASUREMENT OF ARCING TIME IN CIRCUIT BREAKERS electrical phase decoupling in the switch and along the conductors is high, which
reduces interference. However, damping in the line propagation is weak and capacitive
Mike Chapman coupling at bus bars is strong, so interference can be expected. Tests and modelling are
being carried out to analyze how the arc signal of one phase arrives at the other
Goal of the project measurement locations, its significance, and if necessary how this interference can be
In the context of condition-based monitoring, a circuit-breaker monitoring algorithm is minimized or even decoupled.
being developed that provides an improved estimate of contact and nozzle damage.
One critical parameter in this algorithm is the duration of the arc for each switching
operation, that is, the time from contact separation to current extinction. The goal of
project MATiC is to develop a reliable method for measuring the individual arcing times
of high-power switching arcs in circuit breakers in a non-invasive manner. There is
special interest in application of the monitoring algorithm to generator circuit breakers,
which have a high profit potential from condition-based monitoring.
State of the art
It is possible to provide the circuit-breaker monitoring algorithm with an estimate of
arcing time. Given a measurement of the arcing-contact travel curve and the geometry
of the arcing zone, an estimate of the arcing time can be calculated. However, this is an
indirect method of estimating the arcing time, so a direct method of measurement is
sought both in the light and electromagnetic signals proceeding from the arc. Solutions
involving measurement of the light have been rejected because they require access
inside the arcing chamber. The presence of electromagnetic energy in substations has
been studied from the perspective of EMI. More recently, it has received attention in
the design of fault-detection and monitoring systems. However, no acceptable solution
has yet been presented concerning the direct measurement of arcing times. In previous
work on this project, measurements with electrical antennas have indicated that such
an approach is indeed feasible. In the last year the measurement methodology has
been improved, and a solution has been found for measuring the arcing time of a single
breaker chamber. Emphasis in the project has therefore been shifted to the problem of
distinguishing between the arcing signals of a three-phase breaker.
Coupling of the arc signal
Figure 1: Sample demodulated arc signal with breaker current Is and terminal voltage Us
A measured arcing signal is displayed in Figure 1, along with the breaker current and
terminal voltage. The high-frequency signal has been filtered, amplified and Outlook for 2004
demodulated before capture; in this way only the essential signal information is
• Verify measurement results in high-power switching tests at ABB facilities
extracted. Previously, strong bursts of measured energy were observed at key events in
the switching process, but the measured signal was weak between these events. The • Develop a solution for three-phase measurement
improved method of measurement has increased the coupled signal strength Partnerships: ABB Switzerland, ABB Greensburg USA
compared to the previous methods, so that a consistent difference in the signal energy
can be observed throughout the entire arc duration. This important result implies a
significant shift in the arcing time measurement algorithm. If the signal energy can be
observed throughout the arc duration, the algorithm is greatly simplified.

Inasmuch as the duration of an individual arc can be measured, the problem of three-
phase measurements becomes one of measuring in a location where the signals are
decoupled, or possibly of decoupling them by means of the measurement. This problem
is being approached in the context of generator circuit breakers, wherein it is
analagous to partial discharge and very fast transient studies in GIS and generators.
Basic initial modelling of the generator circuit and existing knowledge indicate that

54 55

ARCHITECTURE OF ARTIFICIAL INTELLIGENCE FOR CIRCUIT BREAKERS Identification of circuit breaker characteristics for controlled switching
Circuit breakers show variations in closing time due to variations of temperature, idle
Urs Krüsi time, drive energy and close coil voltage. The degree of the variation is design
dependent. For new circuit breakers these characteristics were determined by
Goal of the project specialised tests to adjust the closing time in order to ensure consistent performance
The overall goal of this project is to develop a circuit breaker incorporating some with controlled switching. Such characteristics are usually unknown for circuit breakers
artificial intelligence. To determine what kind and how much artificial intelligence is that were installed more than 10 years ago. The goal is to monitor the operation of the
required, and which tasks it should perform, are the questions to be answered. circuit breaker during its normal operation and separate the influences of the
In the past year the efforts focused on two subjects: temperature, closing coil voltage, idle time and drive pressure. Controlling the breaker
• Comparison of the calculated restrike probability while de-energizing capacitor without a compensation of these variations will usually not result in an acceptable
banks with laboratory measurements performance. By the potential elimination of pre-insertion resistors controlled
• Identification of circuit breaker dependences on external parameters like switching may significantly cut down maintenance costs [2].
temperature, drive energy, idle time and closing coil voltage for circuit breakers Simulations imply that such identification on a circuit breaker is possible. Laboratory
already installed. tests on a 145 kV SF6 circuit breaker were started to verify the methods.
Restrike probability while de-energizing capacitor banks Outlook for 2004
In cooperation with Per Jonsson from ABB Sweden, it was possible to verify the • Complete the laboratory tests of the identification and evaluate the results
theoretical calculation of the restrike probability. It could be shown that with the help • Publish the methods for the identification of circuit breaker characteristics
of the proposed calculation method the predicted potential for up-rating – use of the
breaker at higher voltages and/or frequencies – was in very good agreement with the References
measured characteristics. The results have been accepted for publication at the CIGRE
Session 2004 [1]. [1] U. Krüsi and P. Jonsson, Increased performance of capacitor bank circuit breaker
by controlled opening, accepted for publication at CIGRE Session 2004, Paris
[2] CIGRE WG A3.07, Controlled Switching of HVAC Circuit-breakers Benefits of
Controlled Switching, to be published on, 2004
closing resistors
Partnerships: ABB Switzerland, ABB Greensburg USA

idle time Controller

drive pressure
? closing coil ∆t
voltage temperature ?
P I(bar)
(s) ∆t ∆t

? ? P (bar)

(V) PTV(bar)

Figure 1: Illustration of the project goal: After monitoring the normal operation of the circuit breaker the
controller identifies the characteristics needed to compensate the variations in closing time due to idle
time, closing coil voltage, temperature and drive pressure for controlled switching

56 57

(PROJECT CONOR) Claudia Roero
Micha Semmler, Claudia Roero, Ueli Straumann, Timm H. Teich, Hans-Jürg Weber Introduction
Overall aim Accepting that water drop deformation in the electric field is involved in generating
To elucidate the mechanisms of tonal noise generation on high voltage lines after tonal emission from overhead high voltage lines, facilities have been set up for detailed
precipitation and to assess methods to reduce such noise emissions. optical investigation of the single drop deformation in DC and AC fields and have
proved their capability to yield significant information on the parameter dependence of
General introduction deformation.
Some high voltage lines emit tonal noise, particularly at twice mains frequency Previous studies showed that with a strongly hydrophilic conductor there is a rapid
(100 Hz), when they are wet. The noise levels occurring may be unacceptable in initial decay of acoustical emission on cessation of rain. For this reason one is still
sensitive locations. looking for a suitable coating which shows a rapid initial water run-off and drying of
the conductors and thus accelerates the recovery of the quiet conditions. Moreover, a
100 Hz oscillation of water droplets on model lines had been convincingly recorded in potentially good coating should display a satisfactory stability of its characteristics as
the laboratory but it has been shown (see subsequent text) that the mechanical regards exposure to sunlight, different temperatures, different atmospheric conditions,
motion on its own cannot quantitatively explain the 100 Hz emission. Nevertheless, but above all with increasing age.
water drop deformation in the electric field and associated Taylor instability can be
instrumental in charge injection into the immediate surroundings of the conductors. Deformation of single drops in a DC and an AC field
This charge injection is subject to present investigations by electrical and optical To elucidate the process involved in the acoustic emission from overhead high voltage
means. lines a DC study of single drops was carried out. This investigation provided details on
the instability of a water drop in the electric field in dependence on drop size, surface
Meteorological conditions – rain intensity and duration, drop sizes, fog or rime – have a condition and field strength. Instability controls the size of surviving drops on the
bearing on sound emission levels. The relationships involved are in the process of being conductor, which contribute to sound emission. A record of all stages of water drop
mapped and should be useful in noise level predictions for the utilities. instability development at up to 10000 frames per second was also done, aiming to
catch step by step the ejection of a fine water jet which follows instability.
A remedy against tonal emission after precipitation is seen in the reduction of the To asses the resonance effect with a single water drop a quick demonstration AC
population of readily deformable water drops on the conductors by encouraging run- investigation was done. In an AC field drop motion is found to be subject to different
off and drying as well as attaining drop shapes less subject to deformation. A oscillation modes and resonances in the frequency range of particular interest (100 Hz)
hydrophilic surface can provide these properties, and various surface preparations have dependent on drop size and electrical field strength. Records of motion sequences for
so far been studied under laboratory conditions. different size drops in an AC field demonstrated that the oscillation of drops of smaller
size is rotationally symmetric, while that of the bigger drops is seesaw.
Partnerships: Buwal Switzerland
EnBW Germany Characterization and measurements of different surfaces and coatings
APG Austria In order to find a coating which shows a hydrophilic behaviour for a reasonable length
Illwerke Austria of time, a number of samples and coatings were obtained from different companies
PSEL Switzerland and institutions.
One set of samples analysed was half coated with photoactive material, annealed at
different temperatures. It was proved that exposure to sunlight of those titania
surfaces reduces contact angle also near to zero, but the effect does not persist long
enough to give a significant noise reduction for more than a few hours after exposure
(see sequence in Fig.1).

It was demonstrated that another coating not based on titania is much more effective,
as the hydrophilic effect appears to persist for six months or more even without any
exposition to ultraviolet light. This coating seems to be potentially a good material to
cover high voltage transmission lines. However, additional studies of the properties of
the coating have to be carried out with respect to temperatures, different atmospheric
conditions (rain, fog, snow, ice), increasing age (contamination, corrosion), etc.

58 59

Al, mean rain, 100 kV



Noise Level @ 100 Hz [dB]

46 untreated (1)

44 hydrophilic(2)
42 hydrophobic (3)

sandblasted (4)
Before sunlight, θ = 57° 0 min after sunlight, θ ~ 0° 40
2 4


0.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0 32.0 36.0 40.0 44.0

Time [min]

Figure 2: Decay of sound level after cessation of rain at t = 8’ for an Al tube with different

Al with hydrophilic coating, mean rain

4 hours after sunlight, θ = 23° 16 hours after sunlight, θ = 40° 20.0

Incremental Noise Level @ 100 Hz [dB]

Figure 1: Behaviour of a 50 µl droplet before and after UV irradiation. 4

To assess dependence of acoustical emissions during and after rain on field strength 3
and on surface properties, a series of surface preparations (untreated, hydrophilic 70 kV (1)

coated, hydrophobic coated and sandblasted) on stainless steel and aluminium have 10.0
80 kV (2)
90 kV (3)
been analyzed. As example the case of the 20 mm diameter Al tube is reported in Fig.2 2
100 kV (4)
to compare the effect of different treatments/coatings with respect to noise emissions. 1
All the results refer to 4 minutes ‘mean rain’, a quantity which corresponds to 5.0

0.39 mm/min at the conductor centre.

Another acoustical measurement of interest is that regarding the variation of the noise
level with respect to different electric field strength values. As examples the case of the 0.0

20 mm diameter Al tube covered with the potentially good hydrophilic coating is 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0

Time [min]
shown (Fig. 3). With this coating it is clear that the decay is rather rapid for all the field
strength values and that the immediate return to the baseline amounts typically to
Figure 3: Hydrophilic conductor with different electrical surface field strengths proportional to rms
5 dB/ min. voltages given (note the shorter time scale).
Outlook for 2004
It has been demonstrated that one of the hydrophilic preparations analyzed could be
potentially a good coating for overhead high voltage transmission lines. This needs
additional investigations with regard to the durability of such coating under different
Facilities for automated recording of water drop populations have to be generated and
to be applied to population changes by variation of conditions.

60 61

TONAL EMISSION AND PRECIPITATION (not shown). The total amount of water is reduced by several parameters: a lower
rainfall (Figure 2a), an increase in the electric field strength, and an increasing
Micha Semmler hydrophilicity. Differences we see between model conductors and real cables are
attributed to an artifact from the measurement and have not been investigated in
Overhead high voltage transmission lines are exposed to precipitation which loads
detail so far.
conductors with a certain amount of water, ice or snow. Subject of this part of CONOR
are tonal emissions during the period of precipitation. Currently the focus is on rainfall
When precipitation begins a distinct noise signal is detected in the 100 Hz third-octave
and possible differences rain and fog give rise to.
band and its harmonics. Interestingly all these frequencies exhibit different
Droplet populations on model conductors dependencies with time as the rain continues. The noise level in the 100 Hz third-
The experimental setup includes model conductors that can be subjected to high octave increases rapidly by several decibel right after the rain has started. It remains
voltage (up to 100 kV) and a sprinkler system with several well-characterized nozzles constant or even decreases slightly in heavy rain, while it continues to increase for
(mean droplet size, rainfall) that produce rain or fog. lower rates of precipitation at a lower noise level (Figure 2b). The trend with regard to
If one knows the droplet population (size distribution) on the surface of the cable and rainfall is non-uniform: the noise levels at moderate rain and dense fog are identical
the tonal emission due to a single droplet one should be able to determine the tonal within error limits. Again, differences between model conductor and cable are likely to
emission from the entire high voltage transmission line. originate from experimental limitations.

with HV
pendant droplets 55

Droplet counts / size class

only heavy rain
10 6

Equivalent film thickness / um

Droplet counts / size class

Droplet counts / size class

Noise level @ 100Hz / dB

with HV
hydrophobic hydrophilic 4 heavy rain
no HV
no HV moderate rain
2 80 45
sessile 1 moderate rain
10 dense fog
2 4 6 60 40
Droplet diameter / mm

pendant 40
+ sessile dense fog 35
(a) 10
0 (b) (a) (b)
Droplet diameter 0 1 2 3 4 5 6 0 200 400 600 800 1000 0 200 400 600 800 1000
Droplet diameter / mm Time / s Time / s
Figure 1: Size distributions. (a) Schematically for (solid line) hydrophobic surfaces and (dashed line) Figure 2: Model conductor (full symbols) and real cable (open symbols), both having an oxidized
hydrophilic surfaces. (b) Droplet population on a model conductor (stainless steel, polished) aluminium surface, with applied high voltage (field strength 20 kV/cm) and at different rainfall.(a) Total
with high voltage (dashed line, field strength 20 kV/cm) and without any voltage (solid line) amount of water (expressed as equivalent film thickness a water film would have if the entire water
at heavy rainfall. Lines indicate log-normal fits to sessile droplets. Inset: Size distributions were distributed homogeneously all over the surface) and (b) noise level in the 100 Hz third-octave band
of pendant droplets, fitted by normal distributions. as a function of time elapsed since precipitation has started.

Droplet populations depend on several parameters from which the most important is Prospects for 2004
probably the nature of the conductor’s surface. A hydrophilic surface leads to an • Physical models that explain the observations have not yet been developed. So far
aqueous film (50-100 µm thick) around the cable and large pendant drops on the the investigations have been descriptive only. It is essential to derive a model that
underside. In contrast one observes a bimodal population if the surface is hydrophobic: is able to predict the size distribution and the total amount of water from any rain
numerous small sessile droplets on top and a second mode of larger sessile and for a given conductor, possibly with a modified surface. Furthermore it is of
pendant droplets (Figure 1a). The presence of an electric field yields an enhanced practical interest to find a threshold precipitation at which noise emissions become
number of small droplets as compared to a population without electric field (Figure 1b). significant.
Changes in the precipitation produce nonuniform trends, investigations of which have
to be continued.
Total amount of water and noise level
Within the first 1000 s a model conductor or real cable is exposed to precipitation the
total amount of water increases monotonously (Figure 2a) except for sandblasted
surfaces, where some reduction in resident water is observed after about 300-500 s

62 63



Ueli Straumann 3
10 -80

Frequency [Hz]
The idea that the 100 Hz tonal emission from high voltage lines is entirely due to the

Sound level [dB]

membrane-like periodic deformation of the water drops, seems to be inadequate. 2
Investigation of the sound level emitted that way from a single drop showed it to be 10 -90
too low. The physical cause of this is generally speaking the multipole nature of such a
source, which is due to the incompressibility of water. Further investigations in the past 10
year showed that a model conductor with mechanical fixed protrusions exposed to
high voltage also produces a tonal emission of 100 Hz. 0
The hypothesis of today suggests that the liquid water drops under high voltage are 10 -4 -3 -2 -110
10 10 10 0 10 20 30
deformed to a more pointed shape. With this form, the enhanced electrical field leads Radius [m] Distance [m]
to discharges, which provide the energy for the tonal emission with a yet unknown Figure 1: The oscillation frequency of sessile water drops against their radii (left side). Calculated sound
process. This hypothesis is backed up by the fact that the discharge current is relatively level against the distance from the source, which consists of the 40 l water drop with oscillation
symmetrical and that the power input of the discharge is more than large enough to amplitude of 40% of its radius (right side).
provide the energy needed for the acoustical emission.
As investigation of the discharge currents pointed out, the discharge current is
Oscillation frequency of drops and acoustical emission relatively symmetric on both half-waves. Partial discharge measurement suggests that
Investigations were made of drops with rotational symmetry and a contact angle of the current is composed of Trichel pulses in the negative half-wave and a current with
90°. The surface R(θ ) of such drops may be expanded in spherical coordinates low dynamics in the positive half-wave. Measurements of the photonic emission give
according to reason to assume that this latter current is a glow discharge. Together with these
discharge currents, there is also the assumption that there are swarms of ions living
⎛ ∞ ~ ⎞ longer than one half-cycle in the air and contributing to the discharge current. This
R(θ ) = Ro + δR(θ ) = Ro ⎜⎜1 + ∑ bl (t )P2l (θ )⎟⎟ ,
⎝ l =1 ⎠ assumption is supported by the values of the transit time, which means the time
~ needed by the heavy ions to cross the discharge gap (see [2]). Calculation of this time
where Ro is the radius of the undeformed drop, P2l (θ ) are standardised Legendre points out that the ions would not cross the entire gap within a half-wave-time.
polynomials and bl (t ) their coefficients.
Following the ideas of [1] and demanding that R(90°) does not change in time, this The energy of the discharge current would be large enough to produce the measured
leads, by taking only the lowest terms as a rough approximation, to the oscillation sound levels.
frequencies shown in Fig. 1. According to Fig. 1, drops with a radius of 1.7 mm or a
volume of 10 µl would have a resonance at 100 Hz. Experiments proved this resonance, 60

but showed also that drops of the size of 40 µl (theoretically 50 Hz oscillation

50 0 kV/cm
frequency) have relatively large oscillation amplitudes. Sound emission of such a source
8.4 kV/cm
is presented in Fig. 1. The emission of this membrane-like, mechanically deformed drop 14.5 kV/cm
seems to be too low: There would be several millions of such sources necessary to

Sound level [dB]

20.7 kV/cm
produce sound levels measured on the 1 m long model conductor.
Acoustical emission from protrusions
The protrusions on the conductor consisted of solder points about 4 mm long. The 20
number of points was about 54. The sound emission of this arrangement is shown in
Fig. 2. 10
Remarkable are the relatively strong tonal emissions of 100 Hz (> 20 dB above the
background) and their higher harmonics (namely 200 and 400 Hz). Thus it seems that 0
there must be an emitting process possible which does not require the periodic





deformation of the drops as acoustical membranes. It seems to be likely that this

















process involves discharge currents. Frequency [Hz]

Figure 2: The sound level measured with the model conductor with solder points.

64 65

Outlook for 2004 Conclusions

Understanding of the discharge current and its link to acoustical emission Water drops reaching instability in the electric field are significant sources for charge
Further modelling of the drop deformation and its influence on the discharge current injection to the surroundings. In the absence of other sharp protrusions, they will be
the only sources at reasonable voltage levels.
[1] Lord Rayleigh, "On the Capillary Phenomena of Jets". Proc. Roy. Soc. London 29, No. All ionisation phases of current pulses are associated with UV light emission, so that
196, 1879. pp. 71-97. optical observation is a valid method to map all ionisation generating processes.
[2] W.L. Lama und C.F. Gallo "Systematic study of the electrical characteristics of the
‘Trichel’ current pulses from negative needle-to-plane coronas".
Journal of Applied Physics, Vol. 45, No. 1, 1974. pp.103-113.


Timm H. Teich, Hans-Jürg Weber
To assist efforts to elucidate the processes involved in the generation of tonal acoustic
emissions from high voltage conductors
A small-scale setup has been used to expose single sessile water drops to globally
homogeneous or moderately inhomogeneous DC electric fields. Discharge currents and
light emission from the arrangement are measured with bandwidths from DC to >
100 MHz. In order to register the passage of small water droplets ejected from drops
reaching instability, a flat band of UV-rich light is passed between the apex of the
sessile drop and the (upper) counter electrode to which the ejected droplet will move
on account of its charge.

Water drops of defined volume (2….100 µl) were placed on the hydrophobic surface of a
metallic hemisphere of 10 mm radius residing centrally on the lower flat electrode of a
30 mm spacing homogeneous field discharge gap. Voltage of either polarity is
increased to first discharge inception. As inception is controlled by drop
deformation/instability, the onset voltage is the same for either polarity or decreases Figure 1: Light emission (top two traces) and current (lower two traces) of Trichel pulse "packets"
with increasing drop size. recorded at discharge inception on a negatively charged 50 µl drop. The "satellite" light pulses (↓ )
signify the passage of a charged droplet towards the anode. The middle two traces are synchronous x20
With negative drops, discharge onset produces a sequence of "packets" of fairly regular expansions made to show that no significant current is associated with those light signals.
Trichel pulse trains [1], each of which is evidently terminated by droplet ejection
(Figure). The "packet" repetition frequency (ca. 20….110 Hz) depends strongly on drop Outlook for 2004
volume and is interpreted here in terms of resonant oscillation of the deformed drops. The validity of the observations presented has to be proved to be maintained in large
At a voltage kept constant the sequence comes to an end when the drop volume has systems.
been sufficiently reduced so that the onset conditions are no longer fulfilled. The fast Single drop coupled discharge and mechanical behaviour has also to be observed under
front of each current pulse representing the ionisation period is associated with UV AC conditions.
emission, while the scattered light signal produced by the passage of the ejected
droplet is – of course – not associated with any appreciable current as the charged References:
droplet motion is too slow to produce a readily measurable current signal. [1] W.L. Lama and C.F. Gallo, "Systematic study of the electrical characteristics of the
With positive drop polarity, behaviour appears to be similar, with burst pulse corona in ‘Trichel’ current pulses from negative needle-to-plane corona",
place of the far more regular but lower amplitude Trichel pulses. J.Appl.Phys. 45(1), 103-113

66 67


95%-confidence intervals

Ruben Vogelsang

Time to breakdown [h]

Goal of the project
The goal of the project is to determine the main factors that lead to reduction of
lifetime in winding insulations of high voltage rotating machines. The results are used
to focus further research on the main factors that reduce lifetime of the insulations in Lower by a Lower by a
Factor 500 ! Factor 23‘300 !
order to improve materials and manufacturing techniques. In the past year, the efforts 1
were focused on determination of time to breakdown of winding insulations with:
• different quality of industrial manufacturing 0.1
• mechanical vibrations between 0 Hz and 50 Hz vibration frequency
Lifetime of winding insulations at different quality of industrial manufacturing 0.01
In previous investigations it has been shown that electrical treeing is the main Material
R1 1,
M1 1,
R2 2,
M2 2,
electrical degradation mechanism of winding insulations [1]. Based on that finding, a manufacturing manufacturing manufacturing manufacturing

new method has been developed to test winding insulations with far less material and Figure 1: Time to breakdown values for two different materials with different types of manufacturing
effort. This method allowed in a short time to investigate different factors that may
influence lifetime of high voltage rotating machines. In order to determine different
quality of industrial manufacturing, the newly developed test arrangement has been 500 µm
used. This arrangement has been described in detail in [1] and [2]. For the tests, two
different types of material were chosen. The winding insulations were all
manufactured in an industrial plant. The bars were taken from the production line. For Voids Delaminations
comparing the results, time to breakdown was also determined for bars of the same of tape layers
materials prepared in a reference process. The reference process is characterised by
500 µm
automatically taping in a special application laboratory. The tests were carried out at a
voltage of 32 kV rms, with a constant insulation thickness of d = 2 mm and at room
Figure 1 shows the 63% values of time to breakdown with the 95% confidence intervals
of the tests with the reference materials and industrially manufactured materials. Figure 2: Material 2 at reference Figure 3: Material 2 at poor industrial
The results show that for tests with both types of manufacturing, time to breakdown is manufacturing manufacturing
significantly lower than for the reference material. The reduction of time to breakdown
Out of the results it can be concluded that manufacturing is a main influence on time
for the 63%-values is by a factor of 500 for material 1 and by factor 23’000 for
to breakdown that determines the potential lifetime of the machines, even before the
material 2! The dramatic difference in time to breakdown shows that the type of
stator bars are inserted into the machines. It shall be mentioned that material 1 and
manufacturing has a significant influence on insulations lifetime, even before the
material 2 were chosen since they are typical materials. The same effect of
winding insulation is installed in the machines.
manufacturing quality on time to breakdown and therefore lifetime may apply for
other materials too. Since testing with the new electrode arrangement demands far
The reason for such a reduction lies in the different material structure. Figure 2 shows
less effort, it is recommended to survey the insulation quality in random tests on
the material at reference production. In contrast, in Figure 3 the material at poor
samples from the production line.
industrial production shows many voids and delaminations at the conductor and
within the insulation itself. Influence of mechanical vibrations on lifetime of winding insulations
For the influence of mechanical load on lifetime of winding insulations, so far only little
When the tapes are not applied tight enough or when the impregnation quality is poor, is known in the literature. For reasons of limitations of test equipment, investigations
the binder resin cannot fill all the regions at the conductor or between the mica tapes were only made at high deflections and low vibration frequencies. In the slot of
well enough. This causes voids and delaminations in the material as seen in Figure 3. rotating machines, there are vibration frequencies of 100 Hz with rather low amplitude.
The voids at the conductor lead to immediate tree inception after voltage application. In order to close that gap, a bending machine was developed that allows testing of the
The delaminations and voids in the insulation lead to very fast tree propagation bars at higher vibration frequencies with deflections of ± 0.25 mm and ± 0.75 mm.
through the material. Both processes caused the very low time to breakdown, shown in Between the deflection of ± 0.25 mm and ± 0.75 mm no difference in time to
Figure 1. breakdown values was found. Figure 4 shows time to breakdown for treeing

68 69

experiments at mechanical vibrations between 0 Hz and 50 Hz and its extrapolation to PDT-COIL – POWER AND DATA TRANSMITTING COMPOSITE COILED TUBING
100 Hz and 120 Hz vibrations.
Stefan Neuhold, Evgeny Murtola
Goal of the project (phase II, demonstration phase)
Research, development and demonstration of an intelligent power and data
transmitting composite coiled tubing for the exploration of hydrocarbons.
In the past year the efforts focused on two subjects:
• Test of prototype conductors with two different insulation materials within the
worst case environmental conditions.
• Computer modelling of the electrical system (frequency converter – PDT-Coil –
drilling motor) with special focus on transient overvoltages and electric losses

Testing of a flat, solid shielded electrical conductor (with chemical protection layer) for
integration into a tube wall
Detailed descriptions of the applications of a flat, solid shielded electrical conductor for
integration into a tube wall has been given in a previous report [1] and paper [2]. In PDT-
COIL phase one, this electrical conductor for integration into a tube wall has been
Figure 4: Time to breakdown for 2 different materials at mechanical vibrations
designed, applied for patenting and manufactured. Two prototype conductors have
The results in Figure 4 show that the vibration frequency has a strong influence on time been produced with different electrical insulation material (1. mica/varnish insulation
to breakdown of the materials. The values are reduced significantly for both materials 2. kapton/teflon insulation). Within the period of this report, these prototype
at vibrations of 50 Hz compared to no vibrations. The extrapolation of the results show conductors have undergone various tests like:
that time to breakdown at 100 Hz vibrations are expected to be 2 h for material A and • Dielectric breakdown tests at different rates of voltage increase
0.1 h for material B. This means a reduction of factor 800 for material A and 850 for • Dielectric loss measurements at 20°C, 155°C and 180°C
material B compared to no vibrations. It can be concluded that mechanical vibrations • Liquid leak rate tests
lead to a significant decrease of time to breakdown of the insulations at higher • Current capacity test of electrical shield
vibration frequencies. Such vibrations must therefore be prevented or reduced in the • Mechanical fatigue life tests (extension and compression up to 2%)
machines. These tests showed that a special focus on the mechanical fatigue life of the insulation
system is necessary. The related stress in service is generated by coiling and uncoiling
Outlook for 2004
of the PDT-COIL tube on a transport reel with 4 metre diameter.
Future studies will be focused to further describe the influence of different
manufacturing types and the influence of temperature on time to breakdown. The System simulation
intention is to determine the main factors that are responsible for the lifetime of All the key system parameters, like conductor impedance or mutual inductance
winding insulations. between different conductors in a COIL are frequency dependent. Along with it,
converter voltage is expected to have broad harmonic content due to use of Pulse
References Width Modulation (PWM). In order to simulate such a system correctly, a model, based
[1] R. Vogelsang, R. Brütsch, K. Fröhlich: "Effect of electrical tree propagation on mica on the Fourier series expansion, has been built using the Matlab/Simulink package. The
insulations", 13th International Symposium on High Voltage Engineering, ISH inverter is simulated by a sine-wave voltage source and its equivalent impedance; COIL
2003, Delft, The Netherlands, August 2003 – by 3 phase line with distributed parameters; motor – by its equivalent impedance and
[2] R. Vogelsang, B. Fruth, K. Fröhlich: "Detection of electrical tree propagation in EMF source.
generator bar insulations by partial discharge measurements", 7th International
Conference on Properties and applications of dielectric materials, ICPADM 2003, Waveforms of the inverter and EMF of the motor are taken from practical experience
Nagoya, Japan, pp. 281 – 285, June 2003 and will be specified more precisely on the basis of data provided by equipment
Partnerships: Gebrüder Meier AG, Regensdorf During simulation, voltages and currents in the system were calculated independently
Ofima/Ofible Power Station AG, Locarno for each frequency within the range of interest, taking into account actual magnitudes
PD-Tech Power Engineering AG, Stetten of parameters. By adding, one can get approximate signals in the time domain.
Von Roll Isola AG, Breitenbach

70 71

Simulations showed possibility of overvoltages and resonances in COIL, considerable MODEL OF A MAGNETICALLY LEVITATED TRAIN
current in the shielding conductor, high power losses.
Peter Kronenberg and Habibo Brechna
Outlook for 2004
• Optimization of the electric insulation system and build-up of a scaled test model Magnetically levitated train systems are expected to play a major role in future
with frequency converter – PDT-COIL – motor and electrical connectors. transportation systems. Actual installations have been built using metallic magnet
coils, with the drawback of high ohmic losses. In this work the potential use a of high
• System simulation: Integration of temperature dependence of parameters and
temperature superconducting (HTS) moving part has been investigated.
specific data from motor/converter supplier; transient simulation and
experimental validation of the model.
An experimental setup (shown below) has been developed. It consists of a cryo-cooled
copper stator winding and an HTS tablet as a mover. Permanent magnets are placed
under the stator winding. The Meissner effect then causes the mover to levitate. The
propulsion of the moving part is caused by the moving sinusoidal field in the air gap
induced by the three phase winding. The stator current is controlled by a PWM
converter and a three phase transformer.

Several HTS materials have been compared for the moving part, with electrical
frequencies ranging from 25 to 400 Hz, thus allowing for speed control. A model for the
displacement of the moving part and the traction force has been established.

The feasibility of a magnetically levitated HTS train system has been confirmed,
although further improvements to the magnetic layout and the converter system will
be required.

Figure 1. Simulink model of system

[1] R. Hug, S.Neuhold. PDT-COIL – Power and data transmitting composite coiled
tubing. Annual Report 2002; Power systems laboratory and high voltage
laboratory; ETH Zürich, pp. 2/40 – 2/41, 2002.
[2] S. Neuhold, K. Fröhlich, D. Inaudi. "Bohrgestänge von der Rolle", SEV BULLETIN
15/2003, pp 32 – 36, 2003.

Partnerships: Airborne Development B.V., Netherlands

Shell Global Solutions International B.V., Netherlands
Aptec B.V., Netherlands
Smartec SA, Switzerland
Katholieke Universiteit Leuven, Belgium

Figure: Setup of the magnetically levitated train model

72 73

3.3 Services offered 4. Publications and Reports

Hans-Jürg Weber 4.1 Publications
Our accredited laboratories provided a wide variety of services for several Swiss
companies and institutions in 2003. Invited Publications
K. Fröhlich, A. Pöltl
Accredited calibration laboratory (SCS 081) "A new Algorithm Enabling Controlled Short Circuit Interruption"
The primary tasks this year were again the calibration of IEEE Transactions on Power Delivery, Vol. 18, No. 3. July 2003
complete Impulse, AC and DC high voltage measuring systems
under operating conditions in the customer's laboratory. K. Fröhlich
Additionally impulse peak voltmeters, standard capacitors, C- "High voltage AC and DC insulation systems – state of the art"
tanδ-bridges and calibrators for PD measuring systems have Proceedings of XIII ISH, International Symposium on High Voltage Engineering, Delft,
been calibrated The Netherlands
26 August 2003, p 421
Accredited testing laboratory (STS 181) K. Fröhlich
Our testing laboratory for tests of electrical properties of "Trends in Technology and Diagnosis of High Voltage Equipment"
components of electrical energy supply performed once more a Proceedings of GCC Power 2003, Muscat, Oman
wide variety of tests according to international standards as well 9 Dec 2003
as following laboratory-developed test procedures.

K. Fröhlich, M. Steurer, Walter Holaus, K. Kaltenegger
"A novel hybrid current-limiting circuit breaker for medium voltage: principle and test
IEEE Transactions on Power Delivery, Vol. 18, No. 2, April 2003, pp 460-467

K. Fröhlich, K. Kaltenegger
"Elektrische Energienetze der Zukunft"
ET Elektrotechnik, Nr 11/2003, pp. 53-56

M. Grader, U. Straumann
"A New Approach to Realize Circuit Breakers with Numerious Series- and parallel-
Connected Low-Current Contacts"
CIGRE International Colloquium, Asset Management of Switching Equipment and New
Trends in Switching Technologies, Sarajevo, Bosnia Herzegovina, 15 - 16 September,
pp 112-117

S. Berger, W. Holaus
"Arcing Contacts and Nozzle Condition Diagnostics by Means of Measure of Thermal
CIGRE Colloquium, Asset Management of Switching Equipment and New Trends in
Switching Technologies, Sarajevo, Bosnia Herzegovina, 15 - 16 September, pp 118-123

74 75

B. Kloeckl, A. Thaler
"Measurements results using the ficticious field current model for PMSM"
5. Presentations
EPE 2003, Toulouse, France, 1 - 5 September
Invited Lectures
B. Kloeckl
"Energiekonversion Downhole – eine Vision" K. Fröhlich
Electrosuisse PES meeting "Geothermal energy conversion", Aarau, Switzerland, 2003 " Introduction to Controlled Switching"
Presented at the CIGRE Workshop Controlled Switch Gear Committee
R. Vogelsang, B. Fruth, K. Fröhlich St. Pete’s Beach, Florida, United States
"Detection of Electrical Tree Propagation in Generator Bar Insulations by Partial 8 May 2003
Discharge Measurements"
Proceedings of the 7th International Conference on Properties and Application of K. Fröhlich
Dielectric Materials, ICPADM 2003, Nagoya, Japan, 1 - 5 June, pp. 281 - 285 "A new Algorithm Enabling Controlled Short Circuit Interruption"
Presented at the IEEE Power Engineering Society General Meeting
R. Vogelsang, R. Brütsch, K. Fröhlich Toronto, Canada
"Effect of electrical tree propagation on breakdown in mica insulations" 15 July 2003
Proceedings of the XIII ISH, International Symposium on High Voltage Engineering,
Delft, NL, 2003, pp. 375 - 378 K. Fröhlich
"High voltage AC and DC insulation systems – State of the art"
R. Vogelsang, R. Brütsch, K. Fröhlich Presented at the XIIIth International Symposium on High Voltage Engineering
"How imperfections in mica insulations influence tree propagation and breakdown Delft, The Netherlands
time" 26 Aug 2003
2003 Annual Report, Conference on Electrical Insulation and Dielectric Phenomena,
CEIDP 2003, Albuquerque, NM, USA, 19 - 22 October, pp. 657 - 660 K. Fröhlich
"Trends in der Technologie von Hochspannungsapparaten und Komponenten"
W. Zaengl Presented at the CIGRE National Committee
"Off-line-Isolationsdiagnostik an Transformatoren" Bern, Switzerland
SGB Transformator-Symposium in Switzerland, Zurich, Switzerland 27 November 2003
8 May 2003
K. Fröhlich
W. Hribernik "Trends in Technology and Diagnosis of High Voltage Equipment"
"Modellbasierte Transformatordiagnose", PSEL Tätigkeitsbericht 2002 Presented at the GCC Power 2003, Conference and Exhibition
Muscat, Oman
St. Neuhold, K. Fröhlich, D. Inaudi 9 December 2003
"Bohrgestänge von der Rolle – Integration der elektrischen Leitung in die
Bulletin SEV/VSE, 15/03, 2003
Other presentations
K.-H. Kim, J.-B. Lee, B. Klöckl, K. Fröhlich
"Wavelet and neuro-fuzzy based fault location for combined overhead/cable B. Klöckl
transmission lines" (working title); to be submitted to IEEE Trans. Power Delivery "Energiekonversion Downhole – eine Vision"
Presented at the ETG Meeting, Brugg, Switzerland
19 March 2003

U. Krüsi
"Controlled switching – suitability check for already installed HVAC circuit breakers"
Presented at the CIGRE Workshop Controlled Switch Gear Committee
St. Pete’s Beach, Florida, United States
8 May 2003

76 77

M. Grader
"A New Approach to Realize Circuit Breakers with numerous Series- and parallel-
6. Conferences and Visits
Connected Low-Current Contacts"
Presented at the CIGRE Colloquium, Asset Management of Switching Equipment and 6.1 Conferences and Workshops
New Trends in Switching Technologies
Sarajevo, Bosnia Herzegovina D. Politano, H.J. Weber
16 September 2003 ETG (Energietechnische Gesellschaft des SEV) Informationstagung "Spannungsfelder
bei den Beschaffungsprozessen"
S. Berger Bern, Switzerland
"Arcing Contacts and Nozzle Condition Diagnostics by Means of Measure of Thermal 30 January 2003
Presented at the CIGRE Colloquium, Asset Management of Switching Equipment and K. Fröhlich, U. Krüsi, W. Hribernik, St. Neuhold, B. Klöckl, D. Politano, M. Chapman, S.
New Trends in Switching Technologies Berger, R. Vogelsang
Sarajevo, Bosnia Herzegovina ETG (Energietechnische Gesellschaft) "Geothermische Energie-Erzeugung: Vision oder
16 September 2003 Realität?" Tagung
Brugg, Switzerland
R. Vogelsang 19 March 2003
"Effect of electrical tree propagation on breakdown in mica insulations"
Presented at the XIII ISH, International Symposium on High Voltage Engineering K. Fröhlich
Delft, The Netherlands CIGRE National Committee
28 August 2003 Lugano, Switzerland
11 April 2003
W. Zaengl
"Off-line-Isolationsdiagnostik an Transformatoren" K. Fröhlich
Presented at the SGB Transformator-Symposium in Switzerland CIGRE Technical Committee
Zurich, Switzerland Rio de Janeiro, Brasil
8 May 2003 24 - 25 April 2003

K. Fröhlich, U. Krüsi
CIGRE A03 Working Group Committee Meeting
St. Pete’s Beach, Florida, United States
6 - 7 May 2003

K. Fröhlich, U.Krüsi
CIGRE Workshop Controlled Switching, IEEE Switch Gear Committee
St. Pete’s Beach, Florida, United States
8 May 2003

W. Hribernik, W. Zaengl
SGB Transformator-Symposium in Switzerland
Zurich, Switzerland
8 May 2003

H.J. Weber
Highvolt-Kolloquium 2003
Dresden, Germany
22 - 23 May 2003

78 79

K. Fröhlich A. Bitschi
EBL – Veranstaltung der Elektrik Basel Land zur geothermischen Energie 1. Fachkongress Geothermischer Strom
Basel, Switzerland Neustadt-Glewe, Germany
17 June 2003 12 - 13 November 2003

R. Vogelsang H.J. Weber

ISOTEC 2003 Metas "Seminar über die Messunsicherheit 2003"
Berlin, Germany Bern, Switzerland
16 - 18 June 2003 18 November 2003

K. Fröhlich K. Fröhlich, U. Krüsi, W. Hribernik, St. Neuhold, E. Murtola, P. Favre-Perrod, R. Vogelsang

IEEE Power Engineering Society General Meeting CIGRE, CIRED, Informationsnachmittag "Neue Trends und Nutzen"
Toronto, Canada Bern, Switzerland
13 - 17 July 2003 27 November 2003

K. Fröhlich K. Fröhlich
XIII ISH, International Symposium on High Voltage Engineering CIGRE GCC Power 2003
Delft, The Netherlands Muscat, Oman
25 - 26 August 2003 8 - 9 December 2003

R. Vogelsang
XIII ISH, International Symposium on High Voltage Engineering 6.2 Events
Delft, The Netherlands
25 - 29 August 2003
Besuch des EKZ und des VA TECH-Konzerns
K. Fröhlich, M. Grader, S. Berger "Aktuelle Forschungsschwerpunkte der Fachgruppe Hochspannungstechnologie"
CIGRE Study Committee A3, Chair Organisation: H.J. Weber
Sarajevo, Bosnia Herzegovina 18 August 2003
15 - 16 September 2003
Alumni Tagung 2003 "Elektrische Energiesysteme – Forschung und Visionen"
K. Fröhlich Organisation: H.J. Weber / R. Vogelsang / D. Politano / B. Klöckl
CIGRE Colloquium, Asset Management of Switching Equipment and New Trends in 9 Mai 2003
Switching Technologies, Chair
Sarajevo, Bosnia Herzegovina Mittelschülerinnentage "Forschung und Visionen in der Hochspannungstechnologie"
17 - 18 September 2003 Organisation: H.J. Weber, H. Kienast
3 and 4 June 2003
B. Klöckl
"Energietechnik für die Zukunft", ETG Tagung 2003
Hamburg, Germany 6.3 Visits
7 - 8 Oct 2003
Prof. Jong-Beom Lee
R. Vogelsang Wongkwang University, South Korea
2003 Conference on Electrical Insulation and Dielectric Phenomena, CEIDP 2003, July 2003
Albuquerque, NM, USA
19 - 22 October 2003

H.J. Weber, M. Hinow, U. Krüsi

FKH- / VSE-Fachtagung "Hochspannungsmesswandler"
Brugg-Windisch, Switzerland
12 November 2003

80 81

Joint Activities

1. Colloquia
Topical Problems of Electric Power Engineering
Aktuelle Probleme der Energietechnik
In collaboration with the "Energietechnische Gesellschaft ETG/SEV"

Modellierung von Preisbildungs-Mechanismen in Elektrizitäts- und CO2 - Märkten

Dr. Jacob Bernasconi
ABB Schweiz AG - Corporate Research, Baden-Dättwil, Schweiz
21 January 2003

Regelenergiemarkt in Deutschland
Dr.-Ing. Michael Ritzau
Büro für Energiewirtschaft und technische Planung GmbH, Aachen
17 June 2003

Voltage Dips - A Major Power-Quality Issue

Prof. Dr. Math Bollen
Chalmers University of Technology, Gothenburg, Sweden
1 July 2003

Fault Location in Combined Transmission Lines with Underground Power Cables and
Reduction Methods for Sheath Circulating Currents
Prof. Dr. Jong-Beom Lee
Department of Electrical, Electronic & Information Engineering,
Wonkwang University, South Korea
15 July 2003

IEC 61850 -
Standardisierter Informationsaustausch in Schaltanlagen und darüber hinaus
Dr. Klaus–Peter Brand
Utility Automation Systems, ABB Schweiz AG
18 November 2003

Zuverlässigkeitsanalyse in der Planung elektrischer Energieübertragungsnetze

Dr. Markus Leuzinger
Busarello + Cott + Partner Inc., Erlenbach
25 November 2003

82 83

2. Vision 'Future Energy Networks'

Patrick Favre-Perrod, Martin Geidl, Bernd Klöckl and Gaudenz Koeppel
Recent developments and scenarios for the future energy supply of industrialised
countries have shown that the actual grid structure needs to be improved. Our energy
system is faced with several challenges:
• Environmental aspects are increasingly considered in the commissioning of new
installations and therefore affect the design process and the technological
development of energy equipment more and more. For electric energy, specific
problems can be identified
- To operate non-dispatchable, renewable energy sources economically, solutions
for energy storage are needed
- As rights of way become harder to obtain and the environmental concern about
overhead lines grows, new solutions for underground energy transmission
should be found.
• The needs of future grid participants (producers and consumers) will continue to
change rapidly in the light of liberalisation processes and increasing demand for
customised power quality
• Large scale use of distributed, dispatchable (small CHP, geothermal,...) and non-
dispatchable generation (wind, solar,...) needs solutions for power infeed into
distribution networks. Furthermore, small scale generators rarely are 50 Hz
• New components such as chemical storages, fuel cells, future transportation
systems (e.g. hybrid vehicles), etc. may use chemical as well as electrical energy.
• The increasing energy demand, especially in very large agglomerations of the
future ('mega-cities') can hardly be covered using actual distribution structures.

Considering the interdisciplinary nature of the above problems, a partnership among

four ETH laboratories has been launched. According to the competences of the
participants, four workpackages have been defined:
• Power systems laboratory (Prof. Dr. G. Andresson): System aspects including
System implications of storage and protection, control and communication
• High voltage laboratory (Prof. Dr. K. Fröhlich): Technological aspects including
energy storage technologies and energy transmission technologies in future
• Power electronics laboratory (Prof. Dr. J. Kolar): Design of converters for storage
devices, power generation and power transmission.
• Automatic control laboratory (Prof. Dr. M. Morari): Control of future energy

In a first step, a greenfield approach has been chosen. For each subtopic, ideal solutions
will be worked out regardless of today’s grid. In further steps, transition scenarios will
of course be considered. The activities will include:
• The formulation of a model for energy flow simulations, implemented for rural,
urban and mega-city structures including the consideration of available and
foreseeable energy technologies.

84 85

• Studying the coherence between (stochastic) production, transmission and storage

of energy.
• Identification of missing technological principles with highly desirable properties in
the hypothetic system.

Figure 1 shows a proposal for a new network structure which will be used in the further
project steps: The network consists of interlinked decentrally controlled, autonomous
clusters or energy nodes. The interlinks are operated according to the needs of the
participants, allowing for the transmission of electrical and chemical energy at
customisable frequencies, voltages, pressures, etc. The emphasis is laid on the hybrid
character which all energy systems already have today and that should be integrated in
the technological functionality and the control strategies of future systems for the sake
of economy and sustainability. The linkage between different forms of energy is
illustrated in the figure. Several basic topology proposals will be assessed using a
combination of technical, economical and environmental criteria.

Figure 1: Vision of a fully decentralised network structure consisting of independent energy nodes

Partnerships: ABB Switzerland, Alstom France, ETH Zürich